Today’s guest is Co-founder and VP of business development at Global Cache. Global Cache makes connectivity products that let programmers like me control and automate pretty much anything using whatever software we like.
In an industry full of propriertary solutions, this is quite the unique approach. That’s why I am really looking forward to learning a lot today from Robin Ford.
Thomas Walter is the Strategic Product Marketing Manager for NEC Display Solutions Europe where he also served as product manager for public display solutions. He’s also served as Product and Sales Manager for Sanyo Sales & Marketing Europe.
Highlights From This Episode
Open Modular Intelligence allows a product to be customised in ways a manufacturer cannot forsee.
The lifetime of a computing device is usually shorter than a displays. Modular solutions address this.
Digital Signage is the most common application of on board computing.
Operational benefits include device control for power saving and using sensors to adjust content based on the current environment.
OPS (Open Pluggable Specification) Slots allow third party computing power and video signal interfaces such as HD-BaseT in a standardised format.
Today’s episode is a little different. Instead of interviewing a person, we review a University project that is changing it’s traditional control systems to Raspberry Pi’s and Raspberry Pi Touchscreens. Joining us from the Brigham Young University OIT AV Services team to discuss this project are Daniel Wells: Director of AV Engineering, Brad Streeter: Chief Engineer and Joe Blodgett: Primary Developer of the open source AV control and management solution.
The project can be found here:
https://github.com/byuoitav
Highlights
You can’t know what kind of interfaces the customer will need. Giving them an API gives them the freedom to adapt the system as their needs change.
Active learning classrooms can have up to 10 displays. Because commodity hardware is so affordable, BYU is able to deploy a touchscreen control systems for almost every video display.
Microservers running in a Docker container could be an interesting way to create cross platform device drivers.
The SALT stack can be used to manage policies and security updates on Linux devices.
The ELK stack can be used to store and visualise data.
Mentions
Daniel Wells, Brad Streeter, Joe Blodgett, BYU, Raspberry Pi, Sony, Epson, ELK, Aurora Reax, C#, .net, Node.js, Golang, Docker, SALT, Linux, JWT, CCUMC, University Of Utah
Transcript...
this is a software defined survival where we explore how software defined systems are changing the business of IT today software defined survival how can we make all of our EP systems accessible through our university API projects just like every other I and I think one of the principles that we’re trying to adhere to that %HESITATION is making this possible as we we kind of refuse to support any technology does that that does not have a network based control air it actually mounts the raspberry pi and here we have all sinned in about one package which send clips into if you will of three young boxer forgetting box with an Ethernet cable running down into it powers it and provide the network connectivity and that’s kind of the core compute power that’s running in the room itself and so it’s no longer cost prohibitive in one of these active classrooms where we’ve got eight strains ten screens whatever we’ve got in there each screen actually has its own dedicated touch panel security we think is actually a better story with this solution it was before because when you’re dealing with a proprietary eighty solution all you’re limited to whatever they fought for it right welcome my name is Patrick Murray and today’s episode of software defined survival will be a little bit different we are not interviewing one person we are kind of interviewing a project and this is F. projects that took place sores on going at a university to change over from traditional AB control systems to using raspberry pies and the raspberry pi touch screen ends the development process involved in all of that and joining me to discuss that or Daniel wells director of AV engineering Brad Streeter chief engineer and Joe Blodgett primary developer of the open source software at Brigham Young University gentleman welcome to the show thank you thank you so I’ll just address all of you at once and feel free to jump in with whoever thinks they get an answer best and my first question is is there anything about that introduction that you’d like to correct or expand upon you know that kind of sums up the project itself %HESITATION pretty pretty cleanly so where did the original idea come from so as university we’ve been %HESITATION really trying to promote inside of our center well first off the audio visual group is part of the office of ID the central office ID for university and therefore our online management goes through the the CIO and see I only really been trying to promote %HESITATION open standards %HESITATION web based controls API’s restful web services eccentric and all of our IT systems across campus and several years ago I I come from the I. T. side of the house and I was brought down to be a part of the eighty group of one of the challenges that we were given was how can we make all of our EP systems accessible through our university API projects just like every other action system we’re trying to build so that was a challenge that was given to us was was basically around what kind of API skin you provide us from the classroom specifically an AV systems in general so that’s kind of what set us down this road was trying to explore how to do that we we we played an experimental little bit was trying to make our existing better that we’re using %HESITATION put a layer on top of it if you will to make these systems expose expose an external API layer and Matt got pretty it was difficult to get a proprietary ecosystem to play nicely in an open world and that we had we had a bunch of other cycles as well around automation and %HESITATION improved %HESITATION deployments %HESITATION better monitoring %HESITATION we wanted to play around with big data and see what kind of analytics we could get out of our systems and based on that analytics will kind of trend analysis and can get a I helping us out decision so we had a lot of these big broad goals up at the only if we came quickly to the solution that the only way we can make this happen is to start building it from scratch your CIO was on board with letting us do that as for a pie was a favorite piece of hardware for many of us down here that we’ve been playing with for the better part of the decade in many ways and some of us and so it was %HESITATION and we had a test now available that had been recently released some we figured we’d start there and see where it took us excellent I’m I think it’s really interesting to note that the decision kind of came from the top %HESITATION so the stakeholders said we want your systems exposed like everything else we’re used to on on our IT at work was there anything in particular that they were looking for or was it just a global sweep of everything that’s on the network we want to be able to speak with or were there really any particular status is reports or functionality that that they were curious about I think from AB side was largely philosophical from that level %HESITATION okay there is a a problem project that I think got some interest in a globally where we have a a campus mobile app that is trying to incorporate a lot of the functionality of university into it so that’s one that was used early on as a justification for getting this sort of thing is that we can build intentionally the control of the classroom equipment into them all out and API’s obviously can help make that happen a lot more clearly but I think I think also the the intent is to expose anything that’s available yeah so that all of our systems any any metric and control points is is available via the API nothing gives us the option of determining which of those %HESITATION we make usually in a philosophical aspects of our CEO was really we don’t know what our customers want to do is for us to build interfaces for them often times %HESITATION seems things that maybe we shouldn’t be assuming and so giving our customers API’s and letting them determine what they want to do with our systems philosophical goal of his and so he basically mandated in many ways it all over I all of our systems have API’s for that reason alone and then we’ll see where we go from there was kind of his stance in many ways yeah that makes a lot of sense when you start I mean database is a perfect example you can’t do analytics without collecting the data first and you have this chicken or the egg what is it that I should be storing well you don’t really know what will be interesting until you start using these systems until some time goes on so I can certainly appreciate that mindset so once this decision was made you mention that you all had some experience with the raspberry pi what were the first steps that I imagine you made some kind of proof of concept to start with can you describe those initial baby steps if you will yeah let me just talk a little bit more about this but as as we were first starting out while we were working on a lot of the the efforts to put that layer on top of other system we had a couple of students working for us that had some interest in taking this little brother Joe was one of them at the time he was one of the students we kind of just said Hey what you guys are playing with this thing over here and see what you come up with maybe you can talk a little bit Joe about that yeah so it was kind of an interesting problem was handed to us for leadership in the broadest down they just kind of sat down I mean when they decide how hard would it be to do this without our traditional Bender is would be possible you’ll figure it out and then they just left us alone for like five or six months build something at work kind of money for a down until something else before down into something else tore down and finally they realize that they’re paying us just how fun deadline on SA some of them by the state and so we have a push to have our was really simple thing was controlling a couple of Sony TV’s just changing input turning them on things like that from a really basic you why and so we you know we got together I think all the others simply for five months from start times that demo time and once by Sony them this is the only TV is one of those was that the Sony TV have a really nice web API in a restful API on them as we need to really nice and easy for us to talk to them went through several iterations of this and then demo did for the site that they call they is that leadership council here in the us immigration policy they would like to log in kind of a great go ahead with it and maybe make it actually work which was really cool sprinting ever since I’ve been pushing and they make it actually all so what does that process look like went once you seen it working had this touch panel you returning a TV on off on off the proof of concept was there and you mentioned the word sprints which maybe not everybody here is familiar with so maybe you could talk a little bit about that development process how you define goals set metrics and really make sure you stay on track and are always moving forward so I think they’re taking it from where Joe just left the story off %HESITATION we kind of did things a little bit backwards you know and we we got a bit a little bit by the the the high fidelity prototypes problem that a lot of developers running to where we had a thing that looks really it looks pretty good at the time and it it did some very basic things and I I think our our upper line management figured it was more complete than it really was and and gave us a deadline to start putting it out on campus actually with a building that was being built a new edition to a building and and baby gave us the mandate that that building was going to have this product in it what we need to do to get there and so we started scrambling about point to to kind of really hash out what pieces needed to be done in order for this to be you know quote unquote finished right to be a product tied staying in and %HESITATION yeah so we started working with the business seems to kind of gather the needs of the building what were the what was the bare minimum that we needed in this product in order to meet the needs and then just started scrambling at that point we haven’t even completed a mounting solution for the raspberry pi a touch screen so we started scrambling a little bit of a prototyping effort there so brats team he’s got several different skill sets within his engineering group and so we had some of this team working on we actually three models the prototype a %HESITATION a mail for this because we couldn’t find anything nobody had made a decent flush mount wall mount if you will for the raspberry pi touch screen so we modeled one we’ve got some three printers leader printing up we tested a mountain once we decided we liked that day was our work was local aluminum manufacturers and all this was in a compressed time cycle trying to make me this deadline of this building and then all the while Joe was starting to scramble to really make sure this product was solid on the back and was going to work reliably was gonna work in the distributed fashion that we wanted it to and %HESITATION yeah so so one of the things is I guess entry is that the initial battle it was it was really basic beginner talking directly to TV and you know we kind of had a bunch of hard coded commands in a little work and we realized that we had to have something a little more generally could handle not only one kind of TV but could have a potentially any TV there was a guy control any projector any video switcher any any there’s a lot of that was was taking it back in and trying to figure out how to make a generalized back and %HESITATION and that took up some time but in all honesty the actual control got so the system we were done fairly soon in the process on that demo was in October I wanna see the guts of the control system were done probably by January and I don’t think we really need major modifications to them since it’s been a lot of the things that make enterprise system and persistence of the monitoring and ointments and you know the support ability on it and that’s still kind of in that today in the configuration of it and I was really what was the big push from there and and that’s in large part still we’re kind of doing it and I think one of the principles that we’re trying to adhere to that %HESITATION is making this possible as we would have refused to support any technology does not that does not have a network based control air preferably a restful control where a lot of proprietary protocols out there that are dies on the TCP layer which we can work with we we’ve got a lot of those in the system right now but we’re trying to in part of our efforts with this open source initiative a little bit is to raise awareness to vendors out there on the value of making a nice clean restful interface is too users like us out there all right so we covered a you mention a few things that I’d like to circle back on of course open source we’re gonna come back to that the whole software design getting that generalized backends but before we get there can you paint a picture for what the hardware looks like you you mentioned there were some issues or there was no mounting solution available so what are what are the components that are involved in this solution arm so what one thing before you jump into this particular solution is to keep my but the the software suite we’ve written is not fundamentally tied to the rest it runs Danny can run anywhere we actually have an instance of it in a W. last that allows us to remote control July is a very nice you know are devices in the room that is consumer you know consumer mail hold it off the shelf and it looks nice and works well but the system the control system itself is not fundamentally tied %HESITATION their usual system we have that you know the control system goes in a room you have just a pie how we you know we we have a caller just a regular seven inch touch screen and there’s our friend the mounting brackets it’s basically like a little thing that sits behind the pilot also the from the wall by one inch made of anodized aluminum got some you know air flow events in there and it looks really nice it is designed to record a bracket basically on the front side of a two or three or four games electrical box and then then then the actual housing bent the high screen %HESITATION connects into on the back of the pie screen them out it actually mounts the raspberry pi and the people we have all stated about one package which send clips into if you will of three young boxer forgetting box with an Ethernet cable running down into it powers it and provides the the network connectivity and that’s kind of the core compute power that’s running in the room itself excellent just four of those not familiar with the rest very pie hat is basically another circuit board that you could kind of clip right onto there’s GPI opens on the raspberry pi and you kind of put another component right on there and it’s called the hat so so these raspberry pi touch panel solution just has one cat cable going to it for network and power correct right and how many of these do you have to pull it our turn we need a little over two hundred there were about two hundred and thirty eight glasses of last night okay so roughly around two forty home deployed across campus right now they’re they’re kind of a core component of of our efforts of doing what we’re calling an active classroom it’s a concept within education for kind of changing the way professors teach and it it’s a ball and so trying to put technology in an active classroom there’s the concept of all spaces and break out space isn’t so trying to put this technology was really valuable to us because the touch screens on the pies are relatively inexpensive to other touch panel solutions and so it’s no longer cost prohibitive in one of these active classrooms where we’ve got eight strains ten screens whatever we’ve got in there each screen actually has its own dedicated touch panel to enable the technology that we want so that’s part of the reason why the numbers grow we’ve got quite a few those going out there but I say probably only forty of those are some of the active classrooms in the other two hundred are kind of stand alone systems spread across different rooms in cap across campus so that sounds like a really good sample set what kind of issues have you had or their heat issues or anything that you would expect so the main the main issue had with the raspberry pi has been that there’s no on board members storage inside the flash card or an SD cards in there as a result SD cards wear out and OS’s are are not coming by way of how many rights they do so we had an issue where we had the kind of the first wave of the SD card do with kind of a middle grade middle lower rate STR can for cost reasons those all I don’t know if it’s so widespread failures failing at about the same time which makes sense we put them in about so we had to replace those we’re trying to make it out to a few different ways to control how much logging we do making parts of the OS read only and is using yeah so you will be using a larger capacity %HESITATION SD card can help with that read right programeri cycle and then we’ve also gone with what they call high endurance carts that seems to help a little bit as well that are more intended for higher I owe applications I mean the nature of a card and you’re gonna have to replace it every you know three to five years someone in the in the context of of cost you know where the the high in and of itself was was a motivation from a cost perspective and we’re seeing that having to buy the more expensive cars it’s kind of also had a little bit but still is considerably less than any of the other options that we come across yeah we’ve gone from a fifteen dollar card to a twenty dollars or so its first system right so it’s still relatively inexpensive compared what other prosperous those are one problem in the big hardware issue I believe had any hardware problems other than because it’s a it’s kind of made for makers right it’s easy to miss can figure it like you know jumpers in the wrong place to forget to put a lead in the right place Sir attached how wrong and I think those are really the only things you have is just times or you know somebody doesn’t connect the power number to the right in and as a result screen is internal medicine you correct jumper works and and yes you I don’t think it’s also harder levels and if you just you we did with what we we we have another application bad that we’re using the same hardware solution for about was our scheduling panels we actually put in a room schedulers on the outside of conference rooms and whatnot we’ve been experimenting with using the same way as he issues with those it was a different technology we worked a lot of that’s obviously the way you write your code is going to impact he significantly and the way we read the code on the eighty control side from day one we’ve been very conscientious around around you cycle heat down we’ve cleaned out the become significantly on the and also we don’t have somebody and the fact that it’s aluminum melts helps with the dissipation it’s got some some cooling vents on the side we we don’t settled we are stick around the build the low fifties high forties audacity centigrade %HESITATION most of our most of these things out there right now actually designing amount when it was first assigned to had a fan mail to you know to put a little you know little twenty five millimeter fan in there we found that it really wasn’t necessary to do much in addition to he sinks so it out a little noise to the room and the noise wasn’t worth benefits so we actually pull those out because he wasn’t sure and you were able to be aware and make those kind of decisions because your collecting that data correct yes right so the raspberry pi as an access to go see what the process can pressure and has little time to censor and there she can gather that information on an automated basis I was just gonna say along the lines of of metrics and data collection we we have a pretty aggressive policy that just about anything we can measure and track and is being sent out to an L. cluster elastic sac which is a big data open source big data solution but we have kind of is our standard here on campus okay so we covered a bunch of stuff I wanna go back to the original question what kind of challenges it sounded %HESITATION that the biggest challenges will be discovered in deployment so if things aren’t hooked up properly you’ll have some issues with the SD card as the technology has a limited number of rights us so it doesn’t matter where that SD card is deployed it will need to be replaced every few years so you’ll need to plan for that you mentioned how the software can affect the longevity in the performance of the hardware I think that’s something interesting that we could dig into a little bit for example I have my my team here created an image for the raspberry pi that’s completely in read only mode so there are no rights and maybe you could talk a little bit about that scheduling panel and those issues there where was heating up due to the software and and about your open source solution and maybe even the decision to make it all open source hello there break it up unless someone else open source in the next question so I think open source %HESITATION it just fits well with the facade of philosophy of our team in general I think %HESITATION I’ve been a big fan of open source for the better part of my life and %HESITATION as such as we’re building teams and and kind of making %HESITATION bring bringing in resources the organization we we try to bring in people that think the same way we do in some ways since I think just in general the open source option for us just made sense and as long as our upper management was okay with that that’s the way we’re gonna go every time but at the same time Wu was resources being limited that’s part of the reason like open source if we can get this to the point where others are interested or we have people contributing you know the the driver database layer which is a micro service layer where where where we write a micro service for each type of hardware you know AB grade at some point we’ve got a lot of contribution of somebody writing a Panasonic line and somebody writing a Sony line somebody writing at some line of control and and and not having to do that all in house we’re obviously focusing on the hardware that we care about but it opens the doors and and and for everybody to be better weight when you deal with an open source model of the other saying the other reason too is there’s just a lot of great solutions out there already that are in the open source world and they’ve been vetted and tried and and you stand up when you’re working in a live action firemen there are a lot of amazing tools out there for management for deployment for locking things down for security and so it just made sense to have you know access all these tools relative for Reno free I’m out there and so yeah open source just kind of where we gravitate toward every time we can gravitate toward every time we can’t help me either and the other thing to consider is that the meeting be AB manufactures in general are behind in this kind of technology so this is an option for us to kind of evangelize the value of the benefits of this type of control and get you know more people than just me why you influencing the manufacturers to to go out and develop API’s further for the hardware and philosophically for me anytime we can legitimize the value of the business value of open source our software feels like a with me so absolutely I’m on board with with all of that to seize outages in it though you mentioned device drivers so that’s going to be a big problem for everyone and it’s could be a main reason why you might pick a traditional control system because of all the device drivers that are available right can you imagine a way to create some kind of repository with out choosing a platform what would you say the creator of houses for you in your house or your device drivers are well yes I mean and then we see without using platform you mean just like the rest Iverson is absent or what we mean right so so so one solution is going to be based on dot net another one’s going to be on no J. S. another one might be written in gold laying like your solution and that’s it it’s a big enough problem to deal with device drivers themselves but to develop them for all of these different platforms to run I think in our opinion that’s world were standardization comes into play right now the only reason we need device driver layer is because we have such disparate control %HESITATION interfaces out there right now and if we you know things like at Aurora is doing with the reacts platform trying to come up with a standardized way for vendors to expose the functionality their devices out to everyone is the only way that that problem is going to be solved for us in the meantime right you know out we’re trying to do something similar with %HESITATION standardize the interface for the micro services for example that we’re creating know we’re doing a micro service per protocol or per hardware tight and we have a standardized interface on the way the bat micro service exposes the functionality to our system and so if you build out that it’s it’s a standard if you will there’s no reason why somebody else’s control solution can’t piggyback off that same standard and throw those micro services up in their environment and use whatever the slayer logic you use they want if you will and that same code that people have written around %HESITATION each hardware types to control that software there’s are the hardware in their systems the concerned about language and platform as far as programming goes I think the fact we’ve each GPS are communication there between the services abstracts along the way and then documentation or containers Asian really fixes how about the other half of it where you can build a docker container that use your you know your driver might research if you will and it again it exposes over even if it’s not a standard API as long as it’s it’s a configurable piece and it means a few basic requirements usual not docker container on whatever host you’re doing it on our case with high and you say it’s you know running on this issue port and here’s how you talked about drive America’s service in its you know someone extracted and then we we constructed house that generalized layer of the EDTA I it will do all that translation for you and call out of reach so far as it being a language using C. sharp or dot net over go over peace be over now it is kind of hopefully we’ve we’ve made out a non issue the interface is the same regardless of the technology used to ride it as long as we’re doing that as restful API later excellent thanks for that inside I really appreciate that because obviously device drivers are a huge part of this puzzle and yeah I think %HESITATION standardization will play a role in solving that and this idea of micro services is something I’m gonna dig into a bit more as soon as I get off this call but until that how about the question on everyone’s minds security what kind of issues concerns yeah prevention how do you manage and deploy these things to make sure that they’re secure so first up front you’re only as secure as the leaders linking your change and as of right now the weakest link is almost always the devices themselves so that the manufacturing of the the projectors for example may communicate over Iraq TCP socket so even if you know I put a password and SO we transmit over the network over in plain text so until some of those things get fixed the system just isn’t isn’t going to be as secure as it should be as far as a rather enticing souls go the nice news is that it’s just a leading server so we use our cell stack it’s a kind of an animal shelter is open source management system for living servers and so it’s it’s really is very simple to say you know I want the system to be locked down each year chords on the show houses around the bases %HESITATION you trust your lectures music journal a few days here I know your bank records your range or whatever you can basically lock down a rather quiet using the same best practices methodologies exist for regional excellence and that’s one of the reasons the other reasons why we’re excited about this security we think is actually a better story with this solution than I was before because when you’re dealing with a proprietary eighty solution all you’re limited to whatever they fought for it right and in today’s world we’re still finding ourselves largely Intel that space in a lot of our ecosystems which anyone who’s delved into the security side of things now is a town that is very dated when it comes to it’s just it’s security features some mildly so yeah I was Joe was saying we haven’t once again going back to the open source source world right %HESITATION lyrics is used in some of the most secure datacenters in the world and there are a lot of great tools around securing it and the salt stacked we’ve chosen its policy based things so you don’t find what you were security website and every device you bring in gets that policy and figures itself to here to it right which is one of the great things about using an open source product like lyrics as the core OS two system I’m beyond that believe guys self musician gas in this kind of standard everything that we have secured me tied into our it’s a it’s a Seattle based its cast token authorization and we used a somewhat tokens for authorization ongoing session management but that’s your older our choices but because we’re using standards we can use really any of the same things Edie’s care any other web server yeah and Dan were using modern open open standards %HESITATION we grow with as they grow so you know as as they change in support newer encryption algorithms because you know as shots got hacked or you know whatever and they moved to the during Christian algorithm we get it it comes along with the new packages that come down and use our Christian things like SSH are built in by default we’ve got secure shell access to all these devices across the board and there’s just a lot of benefits that you can take advantage of going down to stroke excellent I think the big takeaways there are that nothing is really secure out of the box is more a matter of configuration and %HESITATION yeah and to think security and how you’re going to to manage these things absolutely you mentioned earlier about the elk stack can you dig a little deeper into that what that means with these letters E. L. K. mean and how you’re using it in an AV sense so L. stands for elastic search log stashing combina it’s a it’s a it was was big data engine based on the leucine charting engine %HESITATION basically it’s it’s and it is a big data ingestion searching and then visualization pieces arm so anything that happens on any of our devices every time it turns on every time it turns off every five minutes ago based status check I generates an advance and it pushes out of and up into this elastic search cluster and elastic search is a very powerful green language with which you can you know do all sorts of cross Asian and searching and filtering in one hour we use that for a lot of our modern pieces you actually generate alerts based on presence or lack of presence of certain events we use it for keeping track of the state of the rooms you know this room is on the line is seventy three we also use it for you know validating some of our assumptions so one of the success stories that happened before I got down here actually with with right down you and are managing director was these are getting much out of our existing system and we had previously been putting a bill requiring every wrong the matrix kind of supported that wasn’t being used as much as we had thought it was and as a result we kind of stopped doing that a little bit and so we’ve allows us to do things like that where you can you know we have some assumptions that we make and that he loves either through those or you know district so based decisions they have a direct effect affect on on the bottom line right and it takes all the discussion out of the decision because %HESITATION sometimes it’s just a feeling I think I need a Blu ray player right now you know right any any final thoughts no I mean the only if people are interested in and learning a little bit more about it where we we try to make ourselves available to answer questions and to interesting we’ve got an email address they can reach out to us at AV API DB dash dash API at the way you know you yeah we’re just having a good time over here and we don’t mind talking about it if anybody thinks you’re looking at the code that organization name and get out of this be why you all I see a movie on the space is not that is just one more so you can go find organization get help in the now that you brought that up really quickly what what would that process look like I have a raspberry pi my desk I’m looking at the ghetto repository what what are the steps to make this happen if I’m if I’m a Christian on an X. Ekstrom programmer and I don’t know much about this modern programming stuff do I have a chance so %HESITATION as Daniel said in the beginning we haven’t done a great job of packaging this up so I could read also stops actually isn’t that complicated I mean it is silent but really what it is we have an image we build and maintain that does it kind of infrastructure all marketing things are set up automatically but also has all the pieces on it fundamentally need to run the systems you can download that some get out and then really there’s a there’s an example doctor compose file doctor composes it’s a sin tax on a tool for downloading and running a bunch of different docker containers so you have to his example now herbal solemn duty all your basic stuff you need up and running so I don’t do that you start running the containers and then you have to you know configure whatever device you’re controlling in such a way that it has right passwords and things and there is a slight amount of configuration that data base you have to do ask you might is I don’t know if we have a great documentation about it if somebody is interested in this please do reach out and we’d be happy to you know detail exactly what you have to do and and he is set up so I would imagine that you know if if you’re trying to get a really basic system set up in a contact us we can probably work and working with you through to get set up in less than a day for sure probably a couple hours if your engine is the basic system but I think I think that’s one of our goals as an organization is his timing is to make the code more consumable to the community how we’ve done a lot of of customisation that isn’t really easy for everyone to understand so on going I think we’ll see improvements in terms of how that code is is available on consumables for for those that maybe don’t have the expertise to someone like Joe and going through the process of helping someone set this up in their ecosystem and hearing the questions and problems they run into will only help us make this more simple style of anybody’s interest please please reach out we helped ideas while relations were kind of in the process right now helping them get asked you to set up we’d love to help anybody itself excellent I love the attitude of %HESITATION generosity and helpfulness and yeah I sense a little bit of passion for this new approach to things it is exciting and it can be intimidating I think those first few steps you hear a lot of words a lot of acronyms flying around but %HESITATION getting to that first I touchable and and the light goes on once that happens I think the motivation comes by its own so that’s why I’m always looking for those initial steps how do I make something happen anything else any other personal information like to give out you’ve given some contact info anything else you’d like to mention I think I think from the perspective of an ATV engineer so I didn’t grow up in the in the IT world %HESITATION it’s it’s doable so don’t don’t fear the the the language and the acronyms and and everything in it takes a little bit of time but it’s it’s learn a ball and it’s durable and it’s exciting anything else you’d like to mention %HESITATION lasting for a there there is a group of %HESITATION university have media people that I called CCM see they have a conference every year this year see CMC is being held at %HESITATION sister university up the up the road from us a little bit the state universities in the area university of Utah and we’re less than a hour away from where she’s young she’s gonna be also anybody attending that interested in in in visiting with us and maybe possibly see what we’re doing we’re we’re not too far away so I will I will also be presenting at the conference sure we do we are doing a presentation of the conference excellent thank you so much for being on the show gentleman thank you if you or anyone on your staff ever considered themselves just in AV programmer join the club that’s how I used to feel I was just an AMX programmer or just Crestron program or whatever language of your choice is whatever it may be there’s generally this feeling in AV that we’re not capable of using modern programming languages and it simply isn’t true sure there’s a learning curve but once you get through it all other languages become easier to learn and it just expands the amount of options you have when designing a system it’s not an either or decision you don’t say I won’t be using these manufacture tools anymore it’s just you have a broader palate to choose from ends here’s what market day founder of idea box had to say about his experience with the online courses at learn eighty programming dot com you know Patrick it’s funny how the smallest things can sometimes be the star of really big ideas %HESITATION before I took the learn ATV programming dot com courses I was in that Terry I’m only a control system programmer kind of mindset rate %HESITATION when he came to new technologies or current technologies like Java script error or things like that for some reason I thought that was different from what I’m doing and what taking your courses flipped for me was not so much what I learned technically taking the courses it was the mindset of well wait a second I’m already doing ninety nine percent of what some of these most of modern programmers are dealing I just have to learn %HESITATION you know the other one percent and that’s really what I did so it’s really been kind of a big change after taking the course %HESITATION and I would really recommend this course to any integrator not only will obviously help their skill set but more importantly it might change their whole mindset %HESITATION which is more important and and and really show them new opportunities open the door so they kind of see problems through a different lens %HESITATION I gotta tell you one of the biggest changes for me was as soon as I become myself HTML CSS javascript and solve the you guys that I can make with those technologies I just couldn’t sell a %HESITATION Crestron touch him again mark is a great example of somebody who takes new information and really applies it I know that mark still sells a lot of Crestron equipment but for him for his company for his customers for his business he needed a better you why he needed another option for user interface and modern programming allowed him to do that so the question is how can you use modern programming to improve your business please go to learn AV programming dot com and wherever you see a sign up button go ahead and sign up and you’ll get some free information to get a feel of my learning style and what kind of information is available and of course it would be an honor to have you in role in one of our courses and help you upgrade your skills and take this industry to the next level thanks for listening software defined survival I hope you found it useful and maybe inspires you to try out something new this week if you have any questions does software defined survival dot com and click the appropriate I’d love to answer questions on the air and if you’d like to help spread the word please subscribe comment and share it with your
Andrew Page has been working at Cornell University for over 16 years and currently Manager of Integrated Audio and Video Engineering.
He has experience designing and delivering for unified communications, digital signage, video conferencing, web conferencing, content delivery networks, webcasting, and video platform services as well as cloud based solutions.
Highlights From This Episode
AV Control with a Raspberry Pi, Raspberry Pi Touchscreen, HTML5 and NodeRED
Ordering Matrix Switchers from Alibaba
Deciding between Contract Manufacturing and Established AV manufacturers
How to Manage and Scale software defined systems
Enterprise monitoring of AV systems
Mentioned In This Episode
Cornell University, Glyph Technologies, Cisco, Zoom, NodeRED, Javascript, Node.js, NPM, Kramer, PJLink, Amazon Alexa, Revit, Blender, Unreal Engine, David Bianchiardi, Raspberry Pi Foundation, Chromium, Angular, Alibaba, Amazon Alexa, Colin Birney
Transcript...
this is a software defined survival where we explore how software defined systems are changing the business of IT today software defined survive another place or suffer great place hardware is in the control system stack and looking at the ability for source software to replace what currently uses a proprietary technologies for controlling user interface the model of that technology being closed proprietary just doesn’t really won’t really get us there really isn’t we’re having something that is anyone can write notes for publish and that you could be really can connect that is you start trying to it will just have these external pressures to do what’s what’s a real change and what’s just sort of a possibility I think this has enough ten chilling this idea of software eating into the control stack %HESITATION I think this has enough potential to disrupt that it’s something that should be taken seriously and looked into and and understood welcome my name is Patrick Murray and today’s guest has been working at Cornell University for over sixteen years and is currently manager of manager of integrated audio and video engineering he has experience designing and delivering for unified communications digital signage video conferencing web conferencing content delivery networks webcasting and video platform services as well as cloud based solutions in some of our previous conversations we’ve talked about some of the alternative solutions he’s been looking into and I’m really looking forward to hearing more about those experiences from Andrew page Andrew welcome to the show Patrick thank you for having me happy to %HESITATION is there anything about that introduction that you’d like to correct or expand upon no you just reminded me though that it is a little old at this point and could use some updating your linkedin profile he mean yes so tell us how did you get started in AV how did you wind up in this little niche industry allow he goes back to public high school and %HESITATION you know messing around with %HESITATION different theater productions and being involved in the tax side of those and friends who had bands you know kind of getting broken to doing the sound and employing these together and seeing what works and what doesn’t work just having that a technical ability to take things apart and putting together and understand how they’re going to work and then that led to when I went to do my undergrad at college I make majored in television and radio production and that kind of solidified that trajectory for me getting into video production and producing things was in the studio and producing stuff in the field and audio production music recording which then after college turned into a couple years doing live sound freelance which is a great gig in the summer but when winter rolls around there’s not a whole lot to do in a central New York unless you want to go on the road and eventually I got a job at a small company called cliff technologies which maybe some of our listeners remember but it was a hard drive manufacturer they need hard drives and CD burners and take back up a music recording industry and also for the film production industry and they would rat him out these devices and the qualifying with various design workstations and non linear editors as I do tech support for them for a couple years and that was a great game so you right out of college you only ring and you pick it up and be like Hey this is serving Hancock and so your goal has done and we had a great customer base we had Sony pictures and and copy it roses was a customer and just a very wide base sting you just never knew who you’re going to be talking to who is going to be calling for support and %HESITATION that that John lesson for a little while until I finally ended up landing this job here at Cornell and I’ve been working my way up through the ranks started out at the language resource center which was a job where I was creating language tapes and digitizing all material managing mac and PC computer lab satellite operations and some other stuff so that really got me a lot of experience with various systems and managing things to be operational there is a consistent level need meaning that the in the system could be down at the right operational and then John opened up for new venue that was being renovated thirteen hundred seat performing arts hall or now and the technical director position opened up and I after that it was a great job so you got to do sound and video conferencing video streaming involved in a lot of great events a lot of concerts visits and that helped me to really see the whole university okay good insight into how the university operates the various in two plays a gone between departments and and very stakeholders so after two years of that moved into a major role and %HESITATION managing the video engineering team at best fourteen and so we support a video conferencing web conferencing service which is based on Cisco infrastructure and soon as a software service package are we offer a video platform service which is billed account for which drives all the online video that comes through Cornell dot EDU and libraries mass and also here and then we have an exact technical support service which we have staff that well or things like town hall meetings and distance learning classes and an operations we have those and commencement convocation and all those special events to go on and the digital signage service and then finally integrated eighty engineering service which does in the consulting is actually and facilities design build process and we do peer review of some of the larger capital projects that comes through and so that’s been my primary focus for about the past six years and and what we’ve been trying to do there is really trying to migrate away from proprietary technologies and move more towards standards based off more traditional IT technologies we can and we’re also looking at taking the hardware stack and migrating into software where it makes sense and then that software itself moving that software as a service for cloud based services so as many of the audio people listening well here we got Andrew here to us to put a headset on so the audio should be just a little better now what you said really a mouthful there so you you took us through your career path and then you started digging into the meat of what this podcast is about that’s the migration to more I. T. based solutions so I just want to comment quickly about that career trajectory ends you know the more I think about it this some background and live sound the music and you know it comes down to production I think it’s a great training field for for many different fields it’s a it’s a great way to learn about technology especially because you’re interested in it right it’s it’s cool to make music and and to put on a show so you’re interested in it you need to learn about this technology you learn about production just because you know the show must go on you learn about showing up on time everything being in the right time in the right place at the right time I think there’s a lot to be said for being in a in a rock band when her teen ager or at least having hanging out with the crowd that those that of course you could see to many different directions but on the more people that I talk to about this the more I realize there are many different ways you could take those skills to more up I don’t know more grown up type of tempers absolutely yeah really I think it it does teach you about the importance of of as you said showing up on time and the show starts at eight OO five you know there’s no there’s no room for wiggle on that and %HESITATION really kind of will hone your ability to work under deadline and to know what deadlines are important and which ones can slip a little bit yeah all very grown up things but I went when you just making music it’s you know that’s the focus and that’s that’s the motivator so I don’t know it’s an interesting thought I think for a for young people for their own motivation so you talked about this move to away from the hardware stack to the I. T. stack of course there’s a reason why we have all this purpose built equipment floating around you can do what we do you can distribute audio video you couldn’t may control systems in touch panels that worked in real time and show control environment you you just couldn’t do it with software even ten years ago I would say certainly not fifteen or twenty years ago or longer but now it is possible so can you talk a little bit more about what that migration looks like how are you evaluating technologies and equipment what kind of things are you looking at just a tell us about what you’ve been up to sure out so the I think the mom one of the places where we first started to see this about over six years ago was in the video playback we’re we’re adding a DVD player or before that a VCR and then of course more recently Blu ray players into every build everything a standard classroom and as we rolled out the video platform service we began to realize that there are opportunities here to right size the technology that was going into each classroom in that we could start to use these cloud based services and is this the whole idea that there was some sort of link there between the cloud services that were being rolled out and what was being built in the classroom or the conference room was kinda new and we start to think well we don’t really need to put these devices and if we can get people to use video streaming to play back their content and there are some hurdles there like around digitizing various pieces of content but we’re able to overcome a lot of those objections and concerns and people were people professors faculty where we’re starting to do it on their own using YouTube so to then kind of shift them to more institutional tools to do that was fairly easy transition and then we’re able to phase out optical media playback devices in our design build and then as we started to get more into a soft was a service for web conferencing we realize the same thing was going to happen with video codecs and we were able to start to instead of incorporating a poly com or Cisco hardware codec into a build we could then put in something like a body %HESITATION Evey bridge and this was several years ago that we started purposely doing this but what we found over time is that it really wasn’t necessarily replacing all of the hardware codecs south records weren’t necessarily placing all the hardware codecs but that we’re just creating more conferencing a campus or just more video conferencing and web conferencing that was occurring with the prevalence of the software and so in some rooms now we just when it had would not have had video conferencing now we have that via some sort of software codec and the rooms where you’re still doing high and distance learning or there’s some specialized need they’ll still have the hardware codecs so it’s interesting that in that regard the softer hasn’t completely displaced a hardware it’s just created more usage of that type technology and we’re we’re now seeing more I’m sorry now where where we’re starting to look at another place where software group place hardware is in the control system stack and looking at the ability for open source software to replace what we currently uses a very proprietary technologies for controlling user interface so I like the way I think this kind of there’s obviously a cultural shift going on because technology is playing more more of a role in everyone’s lives and I think it was interesting to note that your professors worse doing their own streaming with you too right so that they didn’t even need any kind of a purpose built technology to do it the expectation was just there because there’s these pressures from the outside these services are available to pretty much anyone with an internet connection ends I think that that same idea is available in control or at least that same expectation is there from end users with control risers smart TV’s you could control with Alexa everybody has an apple of course it’s not as integrated as we would like it to be as as professionals in this industry but I think the expectation is also there and %HESITATION I’m starting to feel that pressure as well to to keep up with a great looking user interfaces can you tell us more about what you’ve been doing in in the control space and and what technologies in particular you’re looking at the absolutely and that that consumers ation of IT I think is a big pressure and will continue to be a big pressure and and we’ve you know seen it for years in terms of simplicity of design and costs the idea of what I can do this in my living room lies a cost so much why is it so complicated to do this at my work that’s something we’ve always tried to balance is that walk between using commercial and consumer technologies technologies that are familiar to the end users with building things that we can support at an enterprise scale the the technology stack that most interested in right now for control is software called no dread and it’s something that came out of IBM been open sourced in now as part of the I believe it’s a Java foundation and have to look up the exact name of it and this is an open source platform for the internet of things and the way I sort of see it I think the way that it’s described in in their literature is a way to wire together things with web services and and everything in between and it’s a visual basic programming environment so you have these nodes that you drag out onto a canvas and then you wire the nodes up and a node could do a lot of different things so one node could be TCP IP control and another node could be infrared control and then you could have been no debts just doing some sort of logic for you processing the very signals as a comment or you could have a node that is doing HTML and is driving your user interface it’s all based on very familiar technologies for web developers to build around Java script which is like the most popular programming language if you are out there and so a lot of front end developers can be very familiar with this Java script and then the underlying web server technology is no J. S. which is becoming a very popular web server technology so I’m a big fan of no dread anybody who’s been following my blog in and looking at the learning programming dot com website knows that I think node read up first of all I’m really grateful that it came out because using modern programming can be really intimidating for somebody coming from any V. environment even even if you’re an AMX programmer where you’re used to a script type of a language jumping from that to something object oriented something with so many frameworks and different ways of doing things I think it’s intimidating so I’m grateful that no great is out there because I think it’s the easiest way to dip your toe into the waters of of a more modern programming language because like you were saying you just drag over TCP node putting the I. P. address and the connection is there ends to deploy it it’s one click there’s no compiling uploading you can run it on your laptop you could run it pretty much anywhere on a raspberry pi it’s just really flexible and simple to get started so I’m really excited about it and if you do need to get more complicated like you said you’re you’re just writing Java scripts in in these little there’s that function note where you could write as much Java script as you like so it is really flexible and the underlying technology that you mentioned is no J. S. and that is non blocking any event driven which was really always the problem with using like windows XP for control you you would press a button and then it would block the rest of the program until that function was done so no JSA dresses that and I’ve been deploying it on several projects lately and it works just fine so do you have any more thoughts about node reading where you see it going well one of the the I think the really interesting things is the community around it as well and that folks are publishing their flows and they’re publishing their notes so you can go on to the a node red website and you can search and see if someone has created already a floor note for what you’re trying to do and there is already a few in the AV area published so for instance someone has published a node to interface with Kramer presentation switchers cool and there’s there’s a note for PJ link control of projectors and so you can start to where all the stuff together and then it really allows you to move beyond just the AV system in the room because you can tie and external web services you can tie into the field lighting system it’s very easy to tie into that you can find a Luxo or other voice controls we want and voice control it’s really I think the great thing about it is that it is this open platform and allows you to bring in all these various technologies and and make them work in harmony with each other which I always think it’s sort of the goal of us were trying to put together these AB systems is to make it function is one system and the model of that technology being closed and proprietary just doesn’t really won’t really get us there won’t really get us into the future were having something that is open that anyone can write a note for publish and that you could be really can connect with anything that I think is is is absolutely where we have to go in and where the industry will start trying to which it will just have these external pressures to do that yeah I’m kind of I’m looking forward to that future so this this idea of sharing nodes I think is really powerful no J. S. works with node package manager which is basically an online repository that anybody could submit their modules to and when you use no red you basically go into a search bar and look and just look for something you have no idea that it’s and PM underneath there what do you think it’s going to take for the eighty community to develop the habit of publishing their drivers I think it’s some point you’re gonna you’re gonna lose your competitive edge if you’re not publishing your your driver that eventually it’s going to get to a point where in order for you to sell your product you’re going to have to have that that ability for people to interfacing or hardware with an open source control platform that’s interesting so a lot like it happens today manufacture sometimes we even contract a company like mine to write a question on a Max extra module you see the same thing might happen in this open source area I think eventually that’s we’re going to get to yeah fascinating ends in the modern development web development software development it’s actually a common thing to publish open source software just to get some recognition and and show people what kind of skills you have I wonder if that’ll catch on MTV as well possibly so yes we see the next generation of IT slushy V. enthusiasts there too start on their chops I may be a great place for them to demonstrate what they can do so we’re seeing all the praises of these solutions obviously can run anywhere run out or as very pie for thirty five dollars it’s cheap it’s flexible it’s open source what are some of the cons that you could think of I think just the lack of a commercial support structure knowns commercialize this at this point you can’t pick up the phone and called tech support and have someone to walk you through you know the particular issue that you might be having you have to get online and join the forum and you know really said it Diggin yourself to fix things that are happening so that lack of commercials support offering I think is the is the biggest con at this point and and then of course it’s just not proven you know and we don’t want to be putting technologies into our bills that that don’t have a good strong track record and I feel comfortable with the stability of the hardware and software stack and amusing but it’s gonna take some time I think for others to really developed a comfort level yeah a lot of it is perception %HESITATION there’s been plenty of reports I keep mentioning the raspberry pi but I think electronics are also commoditized nowadays I mean a processor Olympics processors Olympics processor course the electronics have some tolerance is built into a but there’s plenty of reports out there of these devices running twenty four seven for many years as far as the software goes really haven’t seen it fall down or do any weird stuff like you mentions you do have to go to Google or go turn to the web to find some answers sometimes when I do that with proprietary systems is what many times so I personally think a lot of this is just perception and I wonder what it will take to actually change that perception because it’s kind of a a you go first type of a scenario at the moment probably just a few brave souls have to step out there and start using it exactly so are you working on anything interesting at the moment that you’d like to share with us something we’ve been working on that that we’ve that we’ve got a lot of head turns a lot of interest is using virtual reality to mock up buildings and rooms and spaces that we’re doing renovation on of that we’re building for the first time and we’re using in that space we’re using rabbit we’re taking the rabbit models and then bringing them into a blender which is an open source tool and texture rising adding textures to the rabbit models and then from there exporting into unreal engine which is working with an oculus rift headset and we’ve been able to get it to the point where we can we take up a building that we’re working on and we can have a person virtually navigate through the entire building go into a into a classroom and I can see the layout of the classrooms you the sight lines look like can get an idea of the screen sizing and it’s been very interesting in terms of the response that we’ve gotten from folks here in our facilities group and I to groups that build these buildings and do the actual designs are architects and engineering folks have X. expressed a lot of interest in it and it’s very compelling you put on the headset and you are really in that space it’s it’s very you know that they were the resolution the fidelity is obviously not the same as reality but the responsiveness as you turn your head as you look around is enough to create the illusion that you’re in that space and you can really get a sense of the space that you just can’t get that sense from a drawing so we’re really excited to start showing this now to our project stakeholders the folks that are making the decisions on budget making the decision you know what schematic design level about you know do we go with one large classroom into smaller classrooms are how we gonna lay out this office space so we see a lot of potential we just started getting this working about a month ago so we haven’t had a lot of experience under belt yet but it shows a lot of promise and I think another place where software can be used within the design build process itself to bring value to the customers absolutely I had a gentleman named David beyond Chardy on the show a few weeks ago and he uses virtual our reality quite a bit for modeling just like you were talking about and his intention there is to manage the risk in in a building projects because you have all these unknowns you’re you’re putting in this technology that may not be tested war parallel technologies that may need to integrate with each other and you’re not quite sure how old how to look in the ends and using virtual well virtual reality is a way for for you to simulate all those different systems working together so I I I kind of took away from that that you’re using it to make design decisions do you see any value as in in terms of managing risk I think one of the things that came out from some of the engineers that came in to see it but the more the %HESITATION the electrical engineering group is being able to look at the various utilities that are going into a space and how they may or may not interact with each other they may you know we may be designing things to try and live in the same exact physical space so some of that risk which I don’t made may not be the same type of risk that that you know John was talking about but no I I we had we’ve really just started to scratch the surface of what we can use this technology for yeah I’m always looking for practical applications but in the beginning of anything like this it’s it’s you really you don’t know what what you don’t know and now there’s just so much potential that you really do need to just start using it to see to see what comes out of it that’s what that’s what our theory was weird okay let’s just start let’s just build it and see what we can do with it and we knew initially I think are our main objective was to try and make decisions about most distant viewer and screen sizing and we creek quickly found that the resolution of the goggles themselves are gonna limit that you’re not gonna be able to really look at you know such and such point font at such a such distance and so that’s legible because the resolution of the display in your eye isn’t it is not dense enough to be able to make those types of decisions but you can get a general sense of the of the size of the screen and there’s all sorts of other design decisions that you can you can make outside of most distant viewer and one of the things we also found just a point of interest is that it can be kind of nauseating to fly through the spaces and some people are okay with it and some people they’re in there for a few minutes and like okay I gotta take a break and so what we’re doing to try and help with that is we’re gonna set up sort of a guided tour mode for folks so that they can put on the headset and it’ll take them through the model in a more no more controlled fashion and then have a pause point we can pause and you can stop and you can look around and you know this is the the such and such classroom and and when you’re ready to proceed to press button then they proceed on to the next point the tour group so providing a more controlled environment for some folks I think will be really helpful and other people are just gonna want bill slicing the whole model and look at every nook and cranny user interface design you’ve got your advanced user that wants to fly yet on the other one who just wants to be taken by the hand and showing around a bit exactly so let’s shift gears back to this control solution with node read what are your ideas on on employing that in in a practical basis how would you deploy it what kind of hardware would you use what kind of user interfaces do you expect how will they be made will be HTML five what do you what does this solution look like in real life so I’m a proof of concept right now that is a raspberry pi and it’s using the raspberry pi foundation touch screen seven inch into seven inch touchscreen mounted into an articulating case on that will sit sits on your desktop and or table top and that screen will articulate up on up and down %HESITATION and the U. I. itself is chromium which is the underlying technology for chrome the odd that the browser chromium is the version that you get on the raspberry pi for for the next and that of the web page itself is running all HTML five and angular which is web technology that allows users to create single paid web apps in the chromium is running a kiosk mode and it’s set to auto boot into kiosk mode so when the same boots up it launches the operating system and washes chromium goes full screen goes to that local host web page which just what page it’s running that the user interface and so that’s four I think you know somewhere sub hundred dollars you’ve got the touch screen and a control processor that is very stable very usable on the the back end of that you could really have any sort of matrix switch your presentation switcher anything that you can find out what the protocol is that its using and if if if it’s published then you’re all set you’re good to go and I found a company that is a manufacturer in China and through Alibaba was able to order a sample unit and for just a few hundred dollars have a very functional presentation switcher with I think it’s got three HDMI and a mini DisplayPort and and a VGA and then each debased T. out an HDMI out into that’s all completely controllable through the note read rats running through the raspberry pi running no grad and then from there you can do up we have said a voice control using Alexa for knocks it out for voice control if you want to add that later on as well wow that’s a that’s pretty fascinating stuff let let me circle back to beat the beginning I’ve been taking some notes here so we describes with the raspberry pi yeah it’s funny we’ve we’ve stumbled upon basically the same technologies so raspberry pi with the the official raspbian image and chromium running in in in kiosk mode and chromium is the open source version of chrome so I I made an image called controlled by its available on get lab which basically does all that for you it’s got raspbian it’s got that chromium that just boots at start up it’s got the SD card in read only mode to protect a little from power outages and then you were talking about using API’s rights you can control anything with an API and that’s what note read does rights you can pull over HTTP notes TCP nodes and I think the cool thing is you can test all that on your laptop without buying any hardware and then deploy the exact same programming to whatever you’ll be using for a control processor and finally this idea of contract manufacturing I I first heard this from Colin purity and it kind of blew my mind but I guess why not go to Ali Baba in order your matrix switcher arm I’m sure that the latest and greatest technology will always need to be provided by purpose built manufacturer with lots of experience in a support team where do you think the line will be or is there one between how do you make that decision between contract manufacturing and in turning to Europe more traditional AB manufactures so I think we’re it’s gonna be is if you’re building two hundred classrooms you’re going to want to try and do it as cost effectively as possible and that’s where you could turn to contract manufacturing you know you could find something off the shelf that that that that manufacture already makes were you could do if you wanted to do something %HESITATION yam give them a spec and have them build that particular product just for you but yeah I think when you get it scale that’s where it starts to really make sense to start turning towards alternative sources for sourcing the actual underlying matrix switching to the signal distribution hardware fascinating but as as these systems scale the risk will be scaling with it how do you go about managing that so there’s and ed at the note read software itself will run on the raspberry pi but also run on a server so you could have a server environment either running locally within your server farm or within Amazon web services or some other infrastructures a service offering and that can provide oversight for all of your installations and you know be hanging each of your devices whether online could send a notification when a device goes off line allow you to actually control various elements of each of the room like so adjusting labels are color schemes are basic functionality all between the those individual pies at fleet of of control pies that you have to use that term back to some sort of central service to monitor and control them and then there’s also enterprise monitoring technologies out there that you could use to be monitoring all your devices so we have here at Cornell we have a a very standard network topology that we deploy where each classroom or conference room we’ll get a fiber from our backbone into the room and then we put in a edge switch same edge switch that goes into our T. our closets maybe smaller ones like a court instead of a twenty four port but sometimes you do you use a twenty four port for larger rooms and each device is on is on that network we have a navy sub netting procedure in place and then all those devices get monitors to enterprise monitoring and that monitoring that is tied into our enterprise notification system so that if a device goes off line whoever is on call will get a push notification on the phone or to get a phone call or whatever they decide is that they want whatever method they decided that they want to have the communication sent to them and so I think as the stars to scale out you just have to start to build enterprise models for monitoring and for notifications and for maintenance of these devices managing a whole fleet as well of of these types of vices got to think about well how do you manage for more updates how do you know when the next version of no J. S. comes out of next version of node package manager comes out you know how you gonna monitor that and I think having a test environment will be important and then also sort of not pushing out updates across the entire ecosystem at once but maybe having a subset of rooms that you push out updates to monitor for stability and then brought to the larger set of rooms so really starts to become more of an active environment like it becomes a production I. T. environment as you start to go down this path and is it really is it before I think the idea of any system as a kind of built it and then it was there for four five six years whatever and then you pry do a forklift on it or maybe you would use you know you’d be able to do some sort of more precise update something more surgical dented a forklift but generally speaking you call people something put in the new stuff I think now start to look at a more of a continuous update model where you’re continuously updating systems you’re adding features adding new a new technologies and and that’s that’s a different model that’s a different way to start thinking about the EVF a structure and start to think about as warm enterprise IT infrastructure absolutely sounds a sounds very interesting and it sounds like you really get the opportunity to to customize like that last point where you’re continually updating not just the end user experience but also that monitoring ands and %HESITATION notification system you can tell that to what best suits your support team absolutely yeah it because certain agree if you’re supporting using slack you can put your you can push notifications to slack at the using UPS genie you know whatever technology they’re using you can build that and it’s not an integration that’s you know it’s not like you’re just stringing things together haphazardly it can be done in a very robust way yeah so I’m what what do we do as a community to start sharing the solutions with each other because the way it is now you go to Infocom and see what manufacturers come out with now we have this alternative where everybody could kind of make their own solution but I think it would be a bad idea to to not share what we learn with each other what how do you see that playing out I think we just need to start developing and %HESITATION publishing are for you know for working with the note read environment or other open source environment start publishing our work there is I think an opportunity to put together some developer kits and a you know bundled together some of the technologies and distribute those for people who want to actually start experimenting with and and using this stuff and then build some sort of community around that so that we can started knowledge share yeah I’m on board for that so one final question on this point I’m I’m not sure how much your organization works with contractors like %HESITATION independent programmers are even integrators consultants where do you see that traditional a V. integration ecosystem what what’s their role in this new kind of an environment where using open source source software and maybe even contracting contract manufacturing the equipment so I think it definitely the installation side was still going to have a need for people to come and install the technology and pull the wires and and we’re all together make a working system out of it but it may be less and less specifying you know spit the particular manufacture that that integrator is comfortable with that maybe less programming and so more along the lines of justice installation component but I think we’re still a little ways off from that reality and there’s a lot of other things that you to play out before before that happens but it would I think would be good for you know anyone who’s in the space to really start learning about what viable technologies are out there and start preparing their people and start thinking about you know what does it look like that what that new future might look like excellent from any final thoughts and things are changing and then technology is always changing and it’s always hard to keep ahead of stuff and no okay what what’s what’s a real change in what’s just sort of a possibility I think this has enough potential like this idea of software eating into the control stack I think this has enough potential to disrupt that it’s it’s something that should be taken seriously and looked into and and understood and certainly maybe not the whole world will move over that way but I think the to be a significant portion of people that just the value proposition is too great to just that’s it took the cost savings and the ability to manage your your infrastructure at an enterprise level we’ll be too attractive in that you’ll you’ll see a lot of folks move over that way but perhaps not everybody great and you thank you so much for being on the show thank you for having me Patrick I hope you can tell from this episode how excited I am that Andrew and I have stumbled upon the same technologies ends Azinger mentions the biggest drawback is there is no one single place to turn to for support to learn exactly how to set these things up how to create a reliable work flow that you could execute on project after project and that’s why we created the control pie image because why should we all be doing the same steps over and over and that’s available completely free and get help just go to get help dot com and search for control pie it’s also why we made the raspberry pi in AV class available at learn AV programming dot com now Andrew inspired me to make a free version of this course so you can dip your toe in the water and get an idea of what it’s like to work with no dread and the raspberry pi in AV systems so go to learn a few programming dot com and sign up for that free course and you’ll get an idea of what we’re talking about ends here’s what market day founder of idea box had to say about his experience with the online courses at learn eighty programming dot com you know Patrick it’s funny how the smallest things can sometimes be the start of a really big ideas %HESITATION before I took the learn ATV programming dot com courses I was in that Terry I’m only a control system programmer kind of mindset rate %HESITATION when he came to new technologies or current technologies like Java script error or things like that for some reason I thought that was different from what I’m doing and what taking your courses flipped for me was not so much what I learned technically taking the courses it was the mindset of well wait a second I’m already doing ninety nine percent of what some of these most of modern programmers are dealing I just have to learn %HESITATION you know the other one percent and that’s really what I did so it’s really been kind of a big change after taking the course %HESITATION and I would really recommend this course to any integrator not only will obviously help their skill set but more importantly it might change their whole mindset %HESITATION which is more important and and and really show them new opportunities open the door so they kind of see problems through a different lens %HESITATION I gotta tell you one of the biggest changes for me was as soon as I become myself HTML CSS javascript and solve the you eyes that I can make with those technologies I just couldn’t sell a %HESITATION Crestron touch him again mark is a great example of somebody who takes new information and really applies it I know that mark still sells a lot of Crestron equipment but for him for his company for his customers for his business he needed a better you why he needed another option for user interface and modern programming allowed him to do that so the question is how can you use modern programming to improve your business please go to learn AV programming dot com and wherever you see a sign up button go ahead and sign up and you’ll get some free information to get a feel of my learning style and what kind of information is available and of course it would be an honor to have you in role in one of our courses and help you upgrade your skills and take this industry to the next level thanks for listening software defined survival I hope you found it useful and maybe it inspires you to try out something new this week if you have any questions does software defined survival dot com click the appropriate I’d love to answer questions on the air and if you’d like to help spread the word please subscribe comment and share thanks
