NewsroomPress Releases In the News
Learn best practices for rapid & safe dev & testing of healthcare experiences
Rym Badri: Hello and welcome to our webinar How to Supercharge Your Healthcare IOT Product Development. My name is Rym Badri and I will be your host today. Today’s presenters are Manish Mistry, Chief Technology Officer and Kinjan Shah, Head of Innovations Lab at Apexon. As CTO, Manish drives Innovation Labs, Next Generation Technologies, and various practices for our Digital Engineering customers. With over 25 years of experience, his extensive background in delivering and managing all phases of Digital Engineering enables him to create solutions for our customers in the areas of mobility, cloud engineering, AI/ML, and so on. He is also an expert in IoT Development and Wearable Engineering. Kinjan is a Principal Architect and Head of Innovations Lab at Apexon. He is involved in identifying and nurturing ideas and innovations and has developed frameworks and tools that help startups and enterprises rapidly launch new mobility and IoT initiatives with minimal risk. And before, we begin, let me review some housekeeping items. First, this webcast is recorded and will be distributed to you through email to share with your internal teams or watch it again later. Second, your line is currently muted. Third, please feel free to submit any questions during the call using the chat or Q&A function. We will answer all of them towards the end of the presentation. We will do our best to keep this webinar to the 45 minutes time limit. At this time, I’ll turn the presentation over to Manish.
Manish Mistry: Hello, folks, thanks for joining this webinar. I hope everybody’s doing okay by their health. It’s very important at this time of the era. So agenda wise, what we’ll cover today is IoT, it’s a big subject by itself, right? So there’s industrial IoT, there’s a consumer IoT, there’s healthcare centric, and there are variables. Sometimes I wonder is IoT the right thing, because we are in a era where everything is connected. So we’ll talk about how a typical architecture as a landscape, the approach towards, right? Then how do you look at your complete parallel development life cycle, we’ll talk about that. There are certain things which many companies have developed on their own, we also have developed on our own, which gives you a good acceleration to take your IoT products or medical health device products or anything, which are having similar architecture and accelerate them to take it to go to market much faster and of course, much cheaper. We’ll talk about few of the other things which are necessary requirement and then how we expedite using automation, so we’ll talk about that also. So let me see if I can share my screen. All right, I can. All right, so let’s look at the typical landscape or kind of a typical reference architecture, that’s the way we call it, right? We have seen that all the years all the devices have become energized. Many devices are becoming multi-function, but the architecture, if you look at it, how that data is transferring from device is still visualizations and between there are five, six different steps it goes through, which is very, very common. So you may in the process of developing either hardware or a Chrome ware on top of it, which may be communicating or multiple different protocols. It could be Zigby, it could BLE, it could be WIFI. It may be talking to either gateway, which is technically accumulation of information and then passing it to the backend servers which hides all the complexity of communication from IoT device to cloud. So this is a very common architecture used by industrial IoT, at the same it is getting used by many healthcare customer also. There is a very common architecture where devices are communicating directly to the mobile devices. We have our generally or a BLER having a hub, which is very used for home kit kind of things where information from devices goes to the hub and then it transfers from there to other servers. Now what is different here? Actually, what we are seeing is that multiple devices are talking, needs are much higher. We’re collecting tremendous amount of data, either if it’s … Let’s look at the variable where which is catering to healthcare. I mean, you never seen so much of gathering of data, which are your personal, in the sense, you actually, technically, a human is asking for digital twin of itself, that can I see what, how am I functioning myself. So architecture becomes very important that how do you handle this data, right? So most of the common architecture nowadays, we see is that you have your cloud, either AWS, G-Cloud or Google, or it could be anything. They do follow typical life cycle. They do have a microservices oriented APIs. They do cater to the real gold offerings towards analytics, so you can create all the pipelines, syndicate the data, massage the data. How do you bring multiple data together from various system and then you have meaning to your use cases, right? At the end of it, you’ll really want a good data storage where you can store all the raw data as well as modified data. Now from this VC Deck web portal and mobile applications are deriving how do you want to manage the data. How do you want to display the data, at the same time, we see the huge value which everybody’s asking for is the visualizations. So while doing this, it’s very important to remember that there is no typical framework which caters to everything, either from development or automation perspective, right? So people are creating different silos of development, it becomes hard to see to it that they are move faster, rapid development, at the same time, how in your agile process, if I develop something on one piece, how it impacts the other pieces. The end to end life cycle becomes extremely hard to manage, and this is a very important consideration as we move towards, especially on the IOT side of the world as we have more devices. Another rich trend we are seeing is that most of the companies are having portfolio of not just one device, right? They have a roadmap of multiple devices. Now how do you containerize your whole gateway of modality? How do you containerize your cloud modality, right? At the same time, how do you give the visualizations to the users? That also becomes extremely important strategy. What we are seeing is the way we have microservices developed over cloud over the period of time, we are seeing the similar structure developed either at gateway or at the data visualization. They already have micro applications getting developed and this is a very, very … Becoming standardized because you can add and remove different types of applications. You can actually show multiple devices to customers if they have connected to, let’s say your phone, you can see only the devices which they are connected, at the same time you can see the data related to that devices as well. So in a nutshell, end to end topology which is becoming very, very crucial in how you architect your IoT roadmap, how do you visualize your data. Let’s see. Then move to the next slide. All right, it’s taking time. Now, next I want to go … Aw, let’s see, it’s not moving … All right, okay. So what are your typical … I mean everybody who’s been through the PDLC or SDLC life cycle, right? Everybody knows that what is important but I would like to still ponder upon few things which are important here, right. As we move towards more and more complex IoT devices, one of biggest challenge happening is do I have a thick edge or a thin edge? In a nutshell, do I have a lot of processing power on my IoT device itself or can I upload my … I transfer all the raw data and upload all the computation power to either to the gateway or to the cloud. So this is a class C debate which has been happening since multiple years and there is no single answer to it, but we do like to have a really good strategy how you are trying to go after your market. Is it a cost impact is very high? Then you want to have a very, very thin edge. Now there’s too much of competition in industry upfront, then you want to have a thick edge, so I mean, also it becomes complex, is that you want multi-core CPU on the IoT device, you want single core. You want CPU at all or not, multiple things are in place. How many sensors you want, what type of data you want to capture. So strategy becomes very important, at the same time, business alignment becomes very important and we really try to see to it that we want to look at all aspect together before you will start blue printing your device or creating your cloud infrastructure, right? Now as we move towards product development of it, design becomes very important, for example, if it’s a health device, which is a … let’s say for example, branded kind of device, which is becoming very norm nowadays for multiple healthcare based devices. Now have you thought of where it’s going to be, have you thought of what kind of allergic reaction a human will have when they put it on the body. So many, the such criterias goes on and it becomes extremely important to go through each and every part of it. Second piece which is really, really important is algorithm development, right? Algorithm development is the key. Now how do you figure out my algorithm is right or wrong, right? So I mean, we know that most of the use cases which we try to stall are usually done in some different ways, so what we tried to do is what we call them as ground proof validation. So ground proof is a very, very important task, the part of algorithm development. Now what has happened, what we have seen is that while you’re going through this whole cycle of development as well as testing continuous cycle, what we come up with something called a Lab-As-A-Service. We create labs and that’s a right way to structure your whole organization, at the same time, via the whole development cycle around the lab, right, right now it should not be … I mean, many companies try to say there’s a form wear team, there’s a hardware team, there’s a cloud team. In fact, it should be one lab team, right, which is coming together and then trying to solve the the problem. That’s a much better organization structure, as well as development methodology everybody should follow. Now, companion app is a big thing and most of the custom IoT based device, industrial IoT deals with it little differently, but consumer IoT does like to deliver very, very differently. It’s a very UI-centric, very user friendly, at the same time, right data needs to be displayed for the customer. Cloud strategy becomes another big important aspect, at the same time when you are going through the whole development cycle, we highly encourage everyone to go through a lot of proof of concepts, right? I mean, and that’s where the lab plays a huge role. If you’re, all teams are surrounded around labs, it becomes easy to try out new things and we’ve got a try fast, fail fast concept, it becomes very, very important and handy. Many times, when people say, “Do I need to have a separate QA/Automation team?” Actually, to be honest, it is important, right? Because your product’s life cycle validation of every aspect is extremely important. Traditionally, the QA casting was the last piece but that has dramatically changed in last few years and we have brought them along with development life cycle and that’s where the LAAS, logical and Lab-As-A-Service makes a huge difference. You bring together a lab’s … You bring them together, let automation members or QA members do their job and they figure out what is passing, what is failing in different conditions. At the same time development teams are busy writing algorithms, writing companion apps or writing the Cloud based definitions, and go through the whole end to end life cycle, very homogenous way. Of course, as we move towards the majority of the product life cycle, we like to do many different types of testing of whether my cloud can scale or not, my data is accurate or not, can I do some simulated testing to see before I go live? Many different scenarios you want to produce if the communication breaks down, what will happen at certain … So there’s a whole plethora of things you want to do, which may not, we may not have part of during the whole development life cycle. This is where we catch a lot of issues, which could arise because of this. Now during all these life cycle, we may find that there is not one single automation framework which can cater to everything and it’s important to think through the whole automation so that any change happening at one piece of it, which is distribute tech systems, how it can actually impact an other area up to the level where either user centricity or data centricity and see whether it’s working or not. This is very important aspect, to do the whole end frame life cycle testing in all of continuous spaces. Now every move forward, I would like and double click and we tie into this. I would like to bring on Kinjan to take you to the journey how, what are the different activities to be performed.
Kinjan Shah Thanks, Manish. So you know, as Manish mentioned about the product life cycle, like what exactly what different areas that we need to focus on. There are other things that we’ll also focus on, which is like how do you look at your activities, right? This is a huge, like you are at least touching five to seven different tactical technology layers, how exactly you plan your products. So any IoT project we have seen, that at least there are different, five or seven different projects that are internally running, reign and then you have multiple teams talking to each other, working with each other and all that and it’s very important for us to kind of look at those activities from like a very high level and then we drill down to the project level. So that’s what this activity layer is about, that okay, what different from earlier, even in the premier development, we have like three different areas that we are focused on. The first thing is how you are actually looking at your sensory interrupts, how you are collecting the data from the sensors, how you are actually processing all the data. So there are certain algorithms or business project that will help to actually ride on top of Chrome ware, that goes on your device. And then from there, we’ll also help to take care of the communication of those data and communicating with the outside world, from the WS. It could be your hub or mobile application, or directly update with. So that’s one thing, which we call as a product development life cycle or embedded product development life cycle where these three major activities has to be taken cared of. If we divide that further, when we get into the algorithm development there are a lot of different decisions that we have to take care of that what kind of algorithm we need, what data will be what are the type of data that we’ll need for the writing those algorithms and all. And you know, what kind of sensors you need to work with algorithms. Of course, you know, these are all software related decisions and strategic decisions that we’re talking about but we also have to think through the hardware, because usually, it goes hand in hand. So when we are talking about a software requirement, we’ll also look at the hardware requirement, what are other limitations. Usually, softwares are easy to kind of manage but when it comes to the hardware and managing them with hardware is difficult. The major things that we have seen is communication layer, where you have different, different, a lot of different protocols that you need to take care of and what fits best in your requirement is very important. That’s where this entire activity landscape and layout is very important. If we look at the mobile application side and whatever its mobile application or you have a edge hub that is actually taking care of things or you have a gateway that is taking of or your data communication, which we call telemetry data or something like that. How do we actually plan that? What all different ways the device can communicate with outside world or device can be controlled from our outside world, right? So both the things that you have to take care of there. So that’s another piece that we are to architect very well, especially from the communication perspective and the functional aspect perspective on the application side and then, of course, underneath, we’ll also take care of all the analytics that you need to implement because once device is out in the market, we also want to kind of see device is behaving in different conditions and the different user types. Those are the things we must plan and decide, like okay, how exactly we do that. Then of course, the entire cloud management wherein we can go manage the data that we are receiving is a huge data, so if we don’t plan it better, it’s really difficult. So it’s better that we plan it, affirm, then think through how exactly that should be. So architecture is very important order and then of course we also have to look at the visualizing piece because that’s how we can provide the meaningful visualizations to the different user groups. So this is one thing, so you know, if we look at the overall activities and how do we want to kind of strategize these, there are two different life cycles that we are working on, like an ambulant product life cycle and the system development, which is a regular sort of the SGLC we can call it. So this is one landscape, so while we are looking at one perspective where it’s the technique, technology, perspective, there’s another perspective that is important. Yeah, which is how this will work in the operational world, right, so let’s say, which stage we are at for the product development, so let’s say if we are starting from the concept design and prototyping, that’s the first step where you just have an idea of a few paragraphs and then you start with that. What different activities we need to plan and it’s very important. Typically, people say that in an RND, we don’t need to plan things and we always believe that in RND and innovations you have to plan better, otherwise, all your budgets are used and you are not getting anything out of it. So you have to be more cautious when it comes to the innovations and RND, because you know, you have to be very, very cautious about what you want to spend and how you want to spend. And that’s one thing very important, we need to kind of understand what are the architecture that we want to implement. What will be our MDP, what should be our MVP that proves our concept and then how we’ll build on top of it? what we need to take care of now, versus what we can develop in the later stage when we actually get to the product level? So those kind of prototyping and all those activities that we need to do and when it comes to device development, we also need to find out that, okay, what kind of sensors do we need, what kind of hardware we need. As Manish mentioned earlier, that what board CPU, what kind of micro controller we need, how many cores of micro controller we need, why we need that many cores? Because the moment you add up through core on the micro controller, you are also looking at the power, what will be your power requirement and all that. So those kind of things, very important, we need to kind of plan it properly before we actually jump onto to the actually, when developing the prototype. What kind of boards will be easy to avail of, what is the life cycle of those boards? What are the new boards that are coming up, what new protocols or technologies are in use? A lot of those areas, right? When we getting to the next phase, which is your development and integrations, this is where usually we see that there are six or seven … Actually, it’s minimum six to seven projects, simultaneously running and very important for operational team to make sure that all those projects are running in simultaneously and achieving meaningful requirements. And that’s where you need to kind of plan it better because this is not just one project that is gathering everything. So there is Chrome ware development going on, there is interface application development going on. There are hardware side, there are a lot of areas that we need to take care of. Cloud application development is going on to also do the IoT platform integrations. CICD is another thing probably you need to implement because that makes sure that everything that you are developing is going right and all. So those kind of things, you need to plan, plan really, really plan your activities and projects properly. We always seen that instead of developing big projects, all the micro projects, when you’ve developed they help a lot. So let’s say, even if we are developing a mobile application, you make sure that in a mobile application, you keep the functional aspects separate and the US interface aspect separate. So there are two separate projects, you just merge them, and when time comes and pays them. Basically, the idea is you make it as micro project as possible, everything should be micro project. Plan in such a way that they are like … You can thread them properly and then when time comes, you can start all the integrations together. So that’s another area. Since we are talking about medical devices and healthcare devices, we fall under one, at least, one or two compliancies. So verification and validations are very important. So best to make, plan that ahead and weave that into your development life cycle. So there is always a validation, verification or a QA affirmation that is going on during your development life cycle and then there is a separate team who’s only focusing on the verifications, from the compliance perspective. Because we have over a period time, we have seen that just because we have not taken cared of certain things, certain security aspects or certain way of development, you have to redo things. So this is very important. While we are talking about this operational view is all these things are very important and we are actually start, we are actually develop. And then, of course, so that’s where you need to focus on all the validation and the verification. Some of the activities that you have written there, that should start with development on like you are unit based and you’re functionally based and you are a functional automation and you know device integration testing and all that. That should start during development and when it comes to validation and verification stage, it is just kind of verifying everything. So those things needs to be planned really, really properly and the fourth thing is production set up, deployments, how exactly your product will go in the production, how you are tracking your devices out in the market, what kind of implementation you have for the data science piece there, where you are actually tracking all the devices on the ground, how they are behaving, what they are doing and making sure that you are controlling some of the activities that are happening on the device like Firmware update is a very basic thing that you will definitely need But apart from Firmware update, how do you support your customer remotely? So those are the things that when you actually start development, you should start thinking about it. Most of the companies we have seen that after first product launch, they realize, that, okay, “Now, this is something that will be really important for us and we have to look into it,” but the idea is we should look at it from the start. So if we look at the Apexon Readiness now, this is where we come in … Knowing all these different pain points and pains that are happening during the product development life cycle, we decided that, “Well,
let’s build some frameworks that can be handy to our clients and when we are actually getting into the development.” The first thing that we did is, let’s look at the entire mobile application, because this is a big piece. Whenever it comes to interface, the device interface with the outside world or especially, when you are developing a company and mobile applications for your devices, it’s very important that these entire DLE interface is managed well and you have three or four different versions of devices and you want to support all of them when the mobile application … if you do all the development from scratch, it’s very difficult. Instead, if you have some kind of SDKs or accelerator that we call it, a framework, it’s available. It’s very easy to kind of just do the configuration and you are good to go. So that’s one thing that we have done and I’ll talk about it in a bit. The second thing is as we talk about the analytics, right, so if your devices are already in the market and you go back all the devices so we have entire monitoring framework and this is not just a monitoring like a simple analytics, this is more like all your errors that are happening and you know what kind of conditions it is happening and you know all the user journeys, how exactly users are using your device and if you want to kind of understand what functionalities are used most and what functionalities, which kind of conditions, which kind of scenarios on the device that’s showing errors and all that, we can do that. So that’s a monitoring framework, we’ll delve a little bit on that. Labs, other services, which is another area that Manish will definitely talk about where you need the … a test with a human subject validation, clinical trials, and bench marking and all that, how exactly you’ll manage that. So that’s another area that is very important. The Chrome ware messaging protocol is another thing where when you are developing the messaging protocol, a VLE messaging on your device, how do you test it? Thee is no way no can test. People usually use … People who develop that actually to kind of do the testing, so we developed that framework. We’ll talk about that also, and the entire test automation. As Manish mentioned, this is very different automation. This is not a regular mobile application development automation or regular functional automation, this a different automation that you need to take care of. So that’s another area that we’ll talk about. Okay, yeah, I can. Okay. Okay, so if we look at this entire Mobile Application Framework, what was the pain that we are addressing here? The main thing is when you do this development, whatever device you have, so there are a lot of different chips that’s available, I cannot have chips that is actually doing VLE 4.2 and VLE file communication. There are TL chips, as there are … and all of device has chips. So there are a lot of those chips out there they’re not doing anything. They all have their own way of communicating with the devices. Of course, you know, there is a main protocol that they are focusing on, but underneath there are a lot of different changes and those who are developers will definitely understand all different thing they are going through. So what we have done is we developed the entire piece and when the devices are changing their behavior, we’ll also make the changes on the application side. So what we did is we developed the application, entire SDK in such a way that it takes care of all these core functionalities that are coming out of these different devices or all these different microcontroller. So we don’t need to worry about on the mobile application side, it’s coming, it’s really, you just configure a few things on that, like you know what different custom services or what different custom characteristics that you need, you just configure everything and you are ready to go. So that’s one thing that we are doing and why we did it is your total V&V cycle becomes very, very short by implementing this approach and by integrating, because this is already tried and tested SDK that we are learning, so that’s one thing that we did. So this is a very highly well architecture, what it shows and where block that sort of thing that let’s say, you already have a BLE device driver which is already built in the … On top of that, we actually put the entire connection and vetting manager. So what device, how, what time you want the scan, how you want it, you don’t need to write any line of work for that. We just configure everything, you know, this SDK will take care of those things automatically, the entire communication handler is managing, making sure that what different services, what different characteristics you have defined it will just bring up all of them when the SDK is initialized. Configuration manager is taking care of all your configurations and what kind of on let’s say you have a custom services where you are also, you have different data series that you are sending and you just configure that and that will be taken care of, but we already question as far as is actually taking care of all the parsing of the data the way it is coming in, whether the data is human deliver or the data is secured, everything is taken cared of out of the box and then we have the entire logging mechanism, which is very important, that we’ll talk later when we talk about the management framework but this logging is taking care of all the log ins. Then of course, you have a micro application, a functional application, that’s another area that we are working on where you don’t need to kind of develop all different samples, because if you are a device manufacturer, there could be different business models, you have a B2B business model or your B2C business model or in some cases you are business to business to consumers business model, right? So all those business models, how do you want to take care of? So from the same application and from the same core base. That’s another thing we have taken care of. So you know, this is the entire SDK framework that is available for use. So this is one thing that we have actually developed and it’s already in action right now. We are using it for an RF, we are using for DIN, analog devices, deep search, and we have our customers are actually right now using it. This is already available on Android, IOS, and another thing that using this entire framework. Let’s say, you have a parallel development going on, so you have a device development going on and you have your mobile application development going on, to develop this interface nowadays, either you use some kind of ready available simulator with the limited functionality or you use this framework which has exactly the same functionality, that behaves as a peripheral the way your application will behave and it uses that, right? So this is where this entire framework comes very handy when it comes to development and we all see it’s significantly releases the development life cycle like people usually need anywhere between six to eight months to complete that. With this, you can significantly reduce that time actually. Of course, it all depends on the complexity also, but that’s what we have seen over a period of … Let’s talk about a little bit on the Medical Device Management Framework. So what this device management framework does is basically, let’s say you have device is already in the market and you want to remotely monitor all the consumer devices or let’s say your industrial devices or their medical devices, right? You want to monitor what different types of errors are occurring, what kind of warnings the device is throwing, in what different conditions that is happening, right? How exactly my users are actually using what kind of function? Let’s say you are on your device, let’s you have three or four different functionalities, how your user are managing those functionalities? What different users are using, what different region they are using, how it’s secondary? So a lot of different user journeys that you can track, and then you can control remotely, so as I mentioned, right? So let’s say your device is completely configurable, some of the behavior on your device, you can just by adding the configurations, you can … by changing the configuration, you can change the behavior of a device, right? Very simple configurations that you need to kind of manage. So this framework, as is, will bring in all this functionalities together and that’s why we’ve built this framework that it’s very easy to use on the Firmware, on your mobile application and Cloud, all the layer, right? So let me walk you through a very high level, that you have your Edge Devices where you have different interfaces available. We put our libraries on the device where these libraries are actually taking care of most of the metadata that you need to do and during their dialogue menu, you can integrate all the custom logs, right? You have mobile application or hub, whatever you have, we can just integrate our library there and then we have the entire backend piece where it’s a data collection, device control and all that. So those things are already available out of the box. So you can do the event logging, you can do the device, normal device logging or the error warning and all the typical logs that you need. And based on that, you can truly create a customer journey there, okay, you know, I want to … so you can create a customer journey there on the dashboards. I’ll talk about the customer journey in a bit. So what functionalities we are able to capture: Data collection, events and user journey, KPI monitoring. So there are certain KPIs that you want to monitor on your devices and the entire tax dep, you can monitor them. Dashboards are available out of the box and you can customize those dashboards, Data Loader is another thing that you, let’s say, on mobile application and Cloud, we are already integrated the third party like Google Analytics or FireBase or something like that and you want to kind of integrate those data with the framework that’s the functionality of a level and you can integrate those analytics to the framework and of course, you can use the customer analytics. Remote Device Management, is again, from where we get all of that. This is a very highly well operational Operations Dashboard, said how many devices are already in the market? How many, if you have to work three different types of devices, different generations of devices, what are those generation and what are the numbers? Do you, guys, from where’s available, what are the typical errors that you a have received in past seven major 24 hours or whatever? How many battery errors? Because this is very important battery errors, why have I elected the three errors here basically? 90% of the time, because of the critical battery situation, you are getting a lot of different errors and you know, your sensors, starts malfunctioning because of the low current enrolled there. That’s where it comes up. The user journeys are something that you actually look at how different types of journeys that you have, like during the device, if your device is doing some kind of functionalities, what different functionalities are used, where the functionalities are completed, where the functionalities were broken, you can check that, right, and you can see how that has worked and this is Event logs, all Event logs, like you know, whenever there are certain event happen, what was the … These are the different a metadata, what was the battery level, what was the RSS level, what different journeys it has broken. Those events have occurred and all that, so that’s another view that we give you. This is complete User Journey, so let’s say your device connection is one journey, your stimulation ramp, that is the excess stimulation or there are some therapies there your device is taking care of, so those are the kind of different journeys that you can take care of and you can define these journeys, right? You can customize these journeys just by tracking the events between … so starting and then end event and then you can see there, between these events there are some mandatory events that are coming, just put that information and it’s available, right? So, yeah. So that’s about these two things. Manish, I would like to hand over Manish for this. Lab-As-A-Service.
Manish Mistry: Yes, so as been mentioned and as a Lab-As-A-Service or the biggest important piece especially in the healthcare world is like, how do you do human subject validation, right? So here, I’ll quickly brush through this is that we do create labs as required based on the type of equipment, how do you assimilate conditions, et cetera. But what happens is many times it’s important that we actually have a test on human, actual human beings. Now this is important a time of algorithm development, to give you an example, if you let’s say, we have sensor based stuff, which is goes on your body, now body’s tone, which is our skin color, makes a huge difference on the sensors, right? So you have to create how, what are the different ways I’m going to test, how am I going to test at the same time what type of humans I’m going to test and that’s how you’ll actually make your algorithms much stronger. Somehow it doesn’t work. Okay, right. So there’s a whole process in that we define all protocols, we figure out different devices. Along with you, actually, we do comparative analysis on the way as well, and what we call the most referenced devices. We do turns up data collection, right, and on top of that, but then we do a whole visualization, so reporting which helps you make your algorithms much stronger. This whole cycle, as I call it, extremely important and that’s the reason we are very big proponent of creating a lab together, making things happen in parallel, developing parallel, and see to it the whole length in life cycle is managed through this process. So this is just an example how we’ve done it for multiple of our customer, to give you idea that how we go through different cycles and it is important that these are strenuous exercise everyone has to go through. You collect on an average maybe 100 to 200 human beings of different types, before you even say that, “Yes, I am now ready to do a much scope of testing.” Kinjan, do you want to talk about Quality Automation as a last piece? Sure, Manish, sure. So again, this is another piece as we discussed earlier that this automation thing is not a regular automation that you need to take or … so let’s take a scenario here where it will place that … You have multiple devices in the market. You have hub and you have applications and variably and you have different pieces that you want to paste here. How do you paste them? You don’t have all the devices in front of you, right? So you download a simulation device band so this and that, architecture is like … We have a simulation bank over here, but the bank you can just implement all the device simulation and automate that from the automation extension. So your test automation tool is separated but this entire exchange is actually helping you run the test cases on different devices, takes that to the … via hub or via mobile application, whatever you are using and it takes it, it considers the cloud and web application. So it’s a full life cycle of the data, how executive data is flowing and whether your functionality is end to end, and working for fine or not and we are actually kept … this extension is actually capturing this data. So the idea here is you use whatever test tools you are using, use whatever BDD or automation where you are using, this extension will actually intended by, play with your automation tool and execute all your requirements, so you don’t need to change, you don’t need to learn new things, it’s just this extension will take care of everything. So that’s one thing that we have done and … can you move on? Yeah, and this is another thing that we have done, which is Firmware Protocol Verification, so this was very much a need for any devices. So that’s it, you have download the device which is communicating to your mobile applications or outside world on different services or different areas, right? Now in a normal condition, all of those OpCodes, that we call the messaging protocol or OpCodes, right? So all the OpCodes are not being called, so why this framework is required is we paste all those OpCodes, we clear the condition on the device using this framework to actually communicate. We actually paste each and ever OpCode whether the device is functioning directly or not and that’s where this entire framework is really handy for us. So that’s what we have done on this framework, so yeah.
Kinjan Shah: Can we move? I think that’s all I have actually. Manish?
Manish Mistry: So coming back to the net, this is where we would like to close and talk a bit. Is a product strategy exactly how you want your whole EDLC life cycle, what are the needs? At the same time, create the whole development, the environment in such a way that you have a end to end iterative life cycle, in this area where teams are very, very distributed, right? It’s become even more important to see to it that every member knows what’s happening at other places, especially in the initial cycles and then they can distribute the work in different pods, as we call it. Also, don’t forget that testing becomes important life cycle, monitoring becomes the important life cycle. Sometimes certain things are important initially, and certain things are very, very important after you’ve gone to the production. With this, I would like to open up for Question and Answer. Rym, over to you.
Rym Badri: Thank you, so at this time as Manish mentioned, we would like to get your questions answered. As a reminder, you can at any time ask question using the chat or Q&A function at the bottom of your screen. I actually have a question for Manish. It says, “We already have a companion app developed. Do you take over the existing code base to integrate your frameworks?”
Manish Mistry: Yes, this is actually already a classic thing, right? I mean everybody’s tried out a few things but as we move towards a new evaluation of let’s say cloud implication, we call it or you want to add new stuff, if you want to merge multiple device together, you have existing code base. Actually, we take over many of those. We are a big believer of if something is already done, how can we reuse it rather than develop it from scratch, right? If the huge optimization, unfortunately we actually optimize it. If there’s opportunity to make it better, we do that as well. We usually are very, very flexible to try to take over, massage it, make it better and the goal is to see to it that are we moving towards the new generation in a scalable way, more secured way, and much better way.
Rym Badri: Perfect, thank you and we actually have another question here that says, “Is the in house state analytics the same as the combination of ELK-analytics and visualization or something different? Manish or Kinjan, whoever wants to answer.
MKinjan Shah: Yeah, let me take this up, so this is not about ELK’s one step, definitely. Most of the companies where you don’t need too many customization. ELK is the best suite for that, where you know, you just take the data and display that, right, and there is not much customization required. This is the best stack to use, but here it’s more about customizing your data. So you have to really allow users to clear their own journeys, you have to allow them to create their own custom reports and all that. That’s where we need little better tool, and that’s what we have used here. It’s not a completely ELK stack, some part of ELK, yes, we have used, but not everything.
Rym Badri: Perfect, thank you and there seems to be one third question here and maybe probably the last one we have time for for Manish. It says, “We are developing a lifestyle wearable. How can HSV Data Collection and Algorithm development play into that?”
Manish Mistry: And I mentioned, right, HSV or I mean Data Collection is the biggest piece in any healthcare wearable, especially, right? Multiple times, you have subjects, or we call them humans are different, which matters, what type of place where you’re wearable is, that is a huge role. What are the different condition human is in? So for example, if condition, a person may be sitting and then the behavior might change if person is running. The behavior might change, so those conditions also plays a huge role. For a couple of customer, what we have done is that because it was extensive patch, development on the human body, we actually created a humanoid kind of concept where we … it will test out multiple devices in parallel and see in different behavior or position of human in automated way, how will it workout and this is a continuous process as a algorithm fine tuning.
Manish Mistry: So in the case of IoT development, the full journey of collecting data till how we see them, how we perceive them, which is consumed by the human being. That full journey is very, very important from the data processing way and I think, I guess we still creating the full platform on top of it to view them, actually plays a big role.
Rym Badri: Great, thank you. Looking at the time, I think we might have one more question we can squeeze in. This one seems to be for Kinjan. “Are healthcare wearables already in the market? How can we integrate the frameworks discussed, especially mobile app DLE and the device management framework.”
Kinjan Shah: Right, yeah, well, this is good question. This is typical scenario that we always face. Like, okay, we already have a product in the market. We don’t have anything on that product how can we do this? So the best approach for this is this, the libraries that we have, we can actually integrate that in the existing application also and existing facts deck. Of course, it requires one BFU, because we have to integrate the library and we’ll go update that from there to your devices but this is very much achievable and this is always the case for us. Whenever we enter in the engagement, this is always the case, but we are … The product journey has already started and we are evolving in a natural recycle, right? So yes, this is possible and we can do it, yeah.
Rym Badri: Thank you and maybe we can close it off with one last, “How does your framework licensing model work?” Manish, would you like to close it off with this one?
Manish Mistry: Yeah, I mean, generally, is a part of the whole product development life cycle and as part of your customer’s journey, we bring it. We generally don’t charge anything as far as we are part of the whole engagement and it all depends how much of that, we have to go many times we have to customize them, which is a very different activity we work with our customer, but … and like we are actually open. We work with our customer to see to it that journey much faster, much smoother, and much more robust, and that’s our goal. It’s not about to make money out of these frameworks.
Rym Badri: Thank you so much and I believe that’s all the questions we are able to answer at this time. Also be sure to check out our Digital Transformation Channel DTV, that brings in industry experts for thought leadership videos on YouTube. Many thanks to our speakers, Manish and Kinjan and thank you to everyone who joined us today. You will receive an email within the next 24 hours with the slide presentation used today and a link to the webcast replay. If you have any questions, please contact us at firstname.lastname@example.org or feel free to reach out to our presenters directly. Their emails are displayed on your screen right now. Thank you all, enjoy the rest of your day.