In our Podcast series “Sh*t You Wish Your Building Did!”, Memoori explores the intersection between technology and commercial buildings through interesting conversations.

For Episode 29, we were joined by Brahm Lategan, Smart Building Consultant at MiX.

We dived into the world of Commercial IoT Building Data. Why are data integration problems in commercial buildings so persistent? AND how can we fix them? Brahm introduces Unified Namespace (UNS), an architecture from Industrial Automation, as a possible solution. UNS can create a centralized, non-hierarchical structure for real-time data exchange across IT and OT systems.

Don’t forget you can now Subscribe to our Podcasts on Soundcloud, iTunes, YouTube, and Spotify.

Transcript of ‘How Can We Fix Building Data Integration Problems?’

James: Hi Brahm welcome to the podcast. Really pleased to have you here today.

Brahm: Yes excellent thanks Jim thanks for inviting me along.

James: Of course tell us a bit about yourself tell us a little bit more about your work at MiX.

Brahm: Sure um so I’m a smart building consultant here at MiX consultancy based out in London um we work predominantly on um uh commercial real estate bu environment um uh projects uh whether it’s fit out type Works um uh retrofit works.

James: Yeah cool and you mentioned to me before you have a background in Industrial Automation as well.

Brahm: Yeah so my roots uh lie within Industrial Automation um so as a as a wher snapper I qualified in uh uh or did training in electrical and electronic engineering as a generic degree um and then since then um the latter early part of my career uh worked in Industrial Automation so that’s traditional PLC type controls um being the person in the control panels with a screwdriver and a laptop coding um and then just as Times Gone by progressed into um the design side of things um the wind the wind blew blew me in a direction um then into building management systems um and and energy management systems um so probably for the past 15 years or so I’ve been integrating um um as an engineer and then the the past five years I’ve been working on uh solution architecture more of the design side of things um over the deployment and um Enterprise systems so Global Systems.

James: Great, all of which is going to be important for this conversation because we’re going to come on to some more Industrial Automation things in a bit but I wanted first the kind of frame this and you know I’ve in the introduction I’ve talked about this kind of traditional data integration problem in commercial buildings so but I want to hear from you you know what what you think is is a good description of that problem and why is so persistent.

Brahm: yeah so um it’s it’s it’s an evolutionary issue I would say um so so traditionally when we come to data integration in in in in buildings um there’s there’s there’s the layer cake model that’s sort of been followed and that’s the traditional method that’s always sort of been there um and and it’s and it’s born out of just what we needed to achieve from the various systems so you know um the sensor talk to the controller the controller talked to the headend the headend talked to the platforms further upstream and historically that was okay because historically we were only interested in that that little vertical in terms of the platform of the system nowadays as technology progresses as we are wanting to gain more exposure across the various platforms that old method of integration is actually holding us back in terms of data integration right so when you when you when you work your way up through the various layers of integration you’ve got a you’ve got a a a technical debt there I think it’s a good way of describing it and that you’ve got to you’ve got to integrate from point to point to point to point in the end you’ve got a lot of commissioning that needs to happen there’s a lot of validation that needs to happen um and inevitably us as humans we make mistakes and and things get missed and then if you make a mistake at at at the bottom part of the layer as it progresses up this layer and it doesn’t get caught then you’ve got corrupt data that sits at the end and which we we all know as useless.

James: So I’ve heard you describe it as a spaghetti soup of network penetration and endpoint management and issues and you mentioned technical debt already.

Brahm: yeah so what what’s what seems to be happening now is we’ve got um uh take for example electricity meter or a meter we got we got metering system and that that the data is valuable to several platforms and it’s also relevant to to other senses um with with within a building so energy is relevant to occupancy right but energy is also relevant to the um um the controls of the building so so the the heating controls the ventilation controls and so forth so we what what seems to be happening now is we have we’ll have an energy management platform and the occupancy platform and if you want to superimpose or if you want to start putting the data sets across each other you’ve then got to start creating Integrations into each platform, so if you then begin to do that for multiple platforms you then end up with this spaghetti soup you’ve got to then create either vpns you you’ve then got to create um access points within the networks where you’ve got baseboard networks so you’ve got separated networks and this is another important aspect here is the convergence of the various networks um you then end up with this mish mash of access points and it becomes very hard to manage it and then equally if you make a change um say you then make a change in a in a device in the occupancy side things and it you then have to change that through each layer and through each step.

James: Right because I mean APIs are great right but there’s still like a API endpoint management right if you if someone changes an a something about an API then that has to be propagated

Brahm: That’s right yeah that’s it.

James: Yeah good okay I think that you know as a I think a lot of people listening to this will recognize that so you know we’ve discussed this before let’s bring in this uh this unified namespace UNS um and again this is where sort of your experience with the Industrial Automation comes from this is something that’s been born from Industrial Automation could you give us sort of a brief overview of what it is.

Brahm: so the challenges that we’ve got within the built environment around data integration those challenges have existed for a long time within the Industrial Automation space so there’s nothing really new there um obviously within the Industrial Automation space if you talk factory settings for example they there’s a lot more competition and there’s there’s a lot more drive to be more efficient because you know one Vehicle Manufacturer wants to manufacture their vehicle slightly cheaper than the other one so that means they get to sell more Vehicles so from that point of view there’s been a lot more necessity and having to have um um uh data integration methods that are scalable um that avoids this um this the spaghetti soup type integration um it allows for the convergence of the various networks and proper convergence not just connecting networks together and calling it convergence.

James: Okay that’s interesting what would you describe then as as proper convergence as opposed to just you know stuffing these things together.

Brahm: Well the reality is you still need to retain a level of isolation right you still need to um so if you have a critical application let’s let’s take a power management system that’s controlling you know fire pumps this is a critical Network you you you don’t want this network to be hacked and the safest way to do it is to um is to keep it separate and isolated you’ve got corporate networks that do not want to sit within the BMS Realms you’ve got you know landlord netw networks that can’t sit on the corporate networks right but we’ve got valuable data that resides within each of these networks that we want to be able to extract so then that’s something that a unified Nam space helps us to do is by extracting the data you’re converging but we’re still retaining that isolation Okay so that’s that’s really a key part of of of what the UNS will thoughts and and and in the industri short donation side it’s the same challenges you’ve got your um your plant networks you corporate networks um they don’t want to talk to each other the it folks and the OT folks have traditionally always sat within their own camps um yeah that sounds but but ultimately to extract this data we’ve we’ve got to come together we’ we’ve got to we’ve got to converge so the unified namespace is a method of doing it not the only method but it is a method of doing it and scalable method as well.

James: This is being used extensively within Industrial Automation already right this isn’t new particular?

Brahm: The concept of the unified name space or or or has been around since 2005 and it was in it was it was thought of by a chap called Walker reyolds over in the United States um who sat in a in a mine and he had thousand thousands of thousands of sensors and he had thousands of thousands of of of various plant that he had to um um that he had to manage and they were all sitting in their little little isolated silos so he said and he initially did it on a spreadsheet on an Excel Spreadsheet actually um was the first sort of iteration of it so as times progressed since then um Technology’s new platforms have come on the market new protocols not new protocols but other protocols that are now becoming more like for example MQTT are now coming into play and that concept that principle has evolved since then um to now being more um more formula as an architecture and something that we can then take and say actually we can apply this within the built environment to with with success we’ we we we’ve got the architectures that we know work within other Industries um how can we now deploy this Within the built environment to leverage the same and to extract all of this data Mass maintaining the the challenges that we um we’ve got in terms of isolated data silid networks the landlord tenant dichotomy thing going on right um so it was also born out of the same challenges that we’ve got right.

James: And as I understand it it’s underpinned by MQTT.

Brahm: Yeah so uh mqtt plays uh plays quite a big part in in this and that’s what’s important in terms of um the convergence as well so mqtt is quite handy in the sense that you can you can just send information out obviously you do need to create some openness within a network to all to to to allow to to allow you to do that but you can manage that very closely very tightly okay um so hence mqtt you know it really suits this this this architecture um but there are other um um ways of doing it so for example VI OPC it can be done via OPC but then you’ve got to look at your network you got to look at is the network have you got how how many isolated or segregated networks do you have and would that be the appropriate method of doing it so the whole point of of the of the unified name space is that it’s an open architecture okay um and it’s not it’s not a box that you go and buy and screw onto the wall plug it onto your network and say I’ve now got a unified name space it’s an architecture that you’ve got to look at your own environment you got to look at your own building and you’ve got to understand well what data do I need to extract from it and then you’ve got to begin to design the unified namespace so it’s it’s a concept right yeah.

James: Yeah and you mentioned it open it being an open architecture and sort of in practice what does open mean here are we talking open source are we what are we talking.

Brahm: So open in the sense of you have got Choice um of um uh how you deploy the architecture so so the the fundamental principles of the architecture is is that it’s it’s number one it’s EDGE driven as much as possible so we want to be able to push as much of the um intelligence if we can or the compute if we can out onto the edge and by the edge I mean the sensors the devices that set on the field if you’ve got an Endor air quality sensor that’s already capable of communicating via mqtt great you don’t need to have an intermediary box that it’s called to communicate with first you can just have that censor data published automatically up into a unified name space right um so Edge as much as you possibly can then it’s um having an iot platform or an iot platform rather should I say um so the iot platform being something like like a node red um or you know trium for example also have uh similar platforms um that can be used um but you can choose you you are want to say go for the nde red route or do you want to go into a trading route and of course others are available sure as far as the iot platform is concerned then you need to have data persistence so the data persistence can be a Time series type database time scale DB it could be mongod DB if you already have a a Microsoft SQL um uh Enterprise subscription within your business there you go you can use that as your as your data persistence and then you need an ontology right again the ontology is um it’s up to you but the point being is that you’ve got you’ve got the building blocks you’ve got here is what you need to achieve the UNS and within each of the component there are there are choices so you then have to go and decide what’s best for my organization what’s best for my business um and you know how do how how can we architect this to suit us individually right as a builder.

James: So there’s sounds like there’s a lot of flexibility built into it as an architecture.

Brahm: Yeah that’s that’s the point there there’s there’s there’s rigidity in in a sense of well you have to have um you know an ontology you have to have some s of some some persistence you have to have um um a an integration type platform an iot platform whether it’s OPC whether it’s a uh something like node rad um and of course others are available um and and and and to try and push the um uh push the the compute as far out into the edge as we possibly can and what this then creates as well is a a single source of Truth and I think that’s the other important thing to to note here is that the unified namespace is a singular source of truth right so all the devices all the platforms would publish their data to the UNS and multiple consumers can consume the data so rather than having to have to the end or air quality sensor you need to have say three connections one to the to the BMS controls one to the display screen wherever it is you have the M the Indo a quality Center publishes his data to the UNS that’s it and then whatever else needs to consume that data asks for it from the UNS and then it’s then within the UNS that you enforce your ontology and you enforce yourami convention and and the governance of that information yeah yeah and it can it can span it can span across isolated Network so you can have something publish from an isolated Network obviously it’s it has to go out through a firewall but you can you can limit it so it doesn’t go in the other direction and it just says that that input there has changed therefore I’m going to tell the UNS that it’s changed and anything else that needs to know about it can they know about it yeah.

James: Okay you mentioned uh data onology sir so again you saying it’s it’s agnostic so for example if you’re already using haystack tagging or brick schema there is no there’s no requirement from this architecture to to be specific.

Brahm: Exactly that the requirement from the architecture is you must have an ontology but if you are um if there’s already extensive Kobe data for example now Kobe obviously as we know it’s not fully encompassing if you want to have semantic tagging but Kobe might be good enough for you right so you can utilize you know that existing ontology again it’s just stipulating that you must have an ontology and you must have a means of structuring and naming your data yeah and then that’s enforced across um so when you then have a provider so for example you want to deploy a digital twin let’s say you can go to your digital twin provider and you say all right here’s a connection point for you there’s one connection Point here’s the here’s the naming convention here’s the ontology you’ve taken the bulk of the integration work away you know and you you you you’ve already got that there so all that have to do is just connect to your unified namespace server um or or or or environment they already know exactly what it is they um how they’ve got to name things at their end and it’s so much easier to manage so you’ve got only got that one point that you’ve got to manage rather than if it’s a digital twin the digital twin needs to have access to so many different parts of the building and the sensors right yeah

James: yep very interesting so I suppose the question I mean this is we’re I mean OB as I said this is this has come from Auto Industrial Automation but I mean we’re talking about something that’s in inverted Commons quite new for buildings commercial buildings I mean where are we where are we with it at the moment I mean what we’re doing here is I suppose is trying to um you know give it some attention right.

Brahm: Yeah so so there are actually already um and again I’m going to come back to the to the point that that that the UNS is an architecture it’s not a it’s not a product that you buy and you screw on the wall and it’s got to be architected to your own requirements so we already have some very forward thinking developers um um that are that have started deploying their own sort of it looks like Loosely looks like like like a UNS architecture but I think we’ve got another another another uh uh acronym out there called an IDL which is an independent data layer um so people might think and say well actually hold on a minute aren’t you just talking about an IDL well an IDL is is is purely about data storage and it’s about the historic data is my interpretation of an ID but I’m pretty sure others might see this and go no it’s not it’s something else um so so there’s there’s been the IDL um type architecture and I’ve seen that they’ve taken that but they’ve also then included live um uh sort of live information that’s and that’s the thing about a UNS it’s live it’s a snapshot for now right um and then I’ve and then I’ve seen some other systems integrators Master systems integrators they’ve also thought about a similar method of how do we cross the Divide how do we cross the various silos um so they’ve begin to develop their own platforms um um but you know you you’ve got to look at scalability and it’s great what they’ve done but can that One Singular systems integrator scale up for a global client for example yeah um have they got that capacity in them to do it so there are there there’s already um work being done on it so I wouldn’t say it’s brand brand brand new I’m not bringing a Brand New Concept into the equation here um and there’s and and people are already starting to think about this with within the industry yeah um as you say I’m trying to just raise awareness around this and I’m just trying to um highlight this and say well hey industry look this is a way of solving the data integration piece um and ultimately that’s what’s going to help us within the smart building environment to suck you know we want to suck all this data out as much as we possibly can we already understand the value of superimposing all this data we already understand what value that the data is going to give us the problem now is is we’ve got these Legacy and historic methods of doing it which is just not conducive to to you know allow us to get all this data out right yeah.

James: It’s a great summary yeah I um yeah I think that it’s super interesting I’ll put some I’ll make sure I’ll put some links um in the show notes in the description where people can find out more but I guess then you know I mean that I think that’s been a great and you know if they want more information I’m sure people will go and look for it I just think is there is there one thing you want someone to take away from this conversation

Brahm: Yeah definitely um so there’s a um there’s a chat within um the Industrial Automation space called Henrik gon Shi um so I can’t I can’t claim lay claim to this saying at all um but what he um he’s got quite a catchy C phrase that he uses and but I think it’s so relevant and and what he says is and and with with within this old digital transformation world that we are now in for for for buildings and digital buildings I think it’s easy to get wowed by Tech yeah and it’s easy to get wed by um uh by the marketing from from from companies so he he says it quite succinctly and he says um on along the process with your within your digital transformation is you got to think about value don’t think about the tech, then you got to think about people and not the tools and if you set yourself clear targets from the start then you’re setting yourself up for Success when it comes to deploying these platforms.

James: Yeah I like that very nice and a good way to finish thank you good. thanks very much for today. I appreciate your thoughts and uh yeah if anyone wants to um ask you a question get hold of you what’s the best way to do that

Brahm: Certainly you can tag me in the comments um for sure if that’s possible um hop on to our website um by all means you can find me on LinkedIn send me a message.