Podcast: Edge starts to play central role in enterprise 5G
So much to cover in 30 minutes: We’re at the beginning of the centralised role that MEC is going to play in enterprise 5G, as Dheeraj Remella, CPO of VoltDB tells Tech Trends Pod. He deftly handles the loaded question of where to bring edge intelligence. And we’re at another watershed, says analyst, Jim Morrish of Transforma Insights; 2020 is a TIPping point in the drive to OpenRAN. Open architecture is in use at Vodafone, Telefonica, Etisalat, and Japan’s Rakuten, to name just a few telcos. Oh, and it works! And before we share what made us laugh, hear how virtual reality training in the enterprise is faster and cheaper than old school methods as you learn 3 steps to improve your business training.
Podcast transcript:
Jeremy Cowan 0:04
Hi, and welcome to the latest Tech Trends podcast brought to you by The Evolving Enterprise, VanillaPlus.com and IoT Now, I’m Jeremy Cowan, and I’m really glad you’re joining us for today’s sometimes serious, sometimes light-hearted look at digital transformation for enterprises.
Our first guest is Dheeraj Remella, chief product officer at VoltDB, a Massachusetts-based data platform that serves the real time streaming needs of 5G applications. And we’re very proud to say that VoltDB are also the sponsors of today’s Tech Trends podcast. So, Dheeraj welcome, first of all, and secondly, thanks for being our sponsor today.
Dheeraj Remella 0:50
Indeed, it’s our pleasure, Jeremy, and thank you so much for having me over here.
Jeremy Cowan 0:54
Well, we really appreciate it. And our second guest, maybe well known to those who follow the Internet of Things, and more recently, enterprise digital transformation. He is Jim Morrish. And along with Matt Hatton, he co-developed and later sold their successful IoT research firm to the giants Gartner. Now, I’m pleased to say they’ve put the band back together to create Transformer Insights, a research firm that’s already well known to readers of our title, The Evolving Enterprise. They’re helping businesses globally to understand how new technologies will change their markets. So, it’s a warm welcome to you too, Jim.
Jim Morrish 1:35
Thank you, Jeremy. It’s great to be here. And thanks for inviting me to join this podcast.
Jeremy Cowan 1:40
Thanks, Jim. Okay, first, we’re going to share some key stories we’ve seen in the world’s tech news lately. Then Jim and I will be quizzing Dheeraj on the growing importance of network edge intelligence. And finally, in What The Tech? we’ll look at some of the weirder stuff, the stories that have made us smile or maybe just grimace. Okay, first the headlines. Jim, let me start with you. What have you found in the news lately.
Jim Morrish 2:11
Thanks, Jeremy. So I found an interesting story around TIP heralding the global drive towards Open RAN. I better unpick that a bit, I suppose. And so TIP, this is the Telecom Infra Project or Telecom Infrastructure Projects. And really, what these guys do is they’re supporting and promoting open infrastructure technology in telecoms networks. And what that means is it effectively allows telecom network software to run on generic hardware. And what that does is means that you don’t have to have your traditional vendor approach of selling integrated hardware plus software, you can have somebody selling software deployed on generic hardware. And there seems to be some significant traction in this space. So the latest update was back in February, and open architecture was in use at divisions in Vodafone, Telefonica, Etisalat, Japan’s Rakuten, and also a Chilean operator, Mundo. So it’s getting some traction and with some big names as well. And really, what I wanted to do is highlight that because it plays into an interesting area. And there’s a huge amount going on, around telecoms networks, particularly around 5G, particularly impacting telecom network provider dynamics. You’ve got things like multi-access edge computing, you’ve got hyperscalers getting involved in the space. So the likes of Microsoft acquiring Affirmed Networks, which is a cloud core proposition. So you can have a mobile network on the cloud deployed, for instance, on AWS, or Microsoft’s Cloud, actually supporting telecom networks.
And this stuff’s has been tried by Vodafone. So again, yeah, it works. But what you’ve got is you’ve got hyperscalers moving into supporting networks, you’ve got the supply dynamics of the software and the hardware for the networks unhinging and separating. You’ve got some geopolitical shenanigans as well, which effectively is opening the doors for some potential new providers of software, you know, the likes of Fujitsu and Samsung and NEC and Mavenir and Parallel Wireless. These guys can offer their software to be deployed on generic hardware, and of course, because it’s just offering software, they can scale that very quickly, so they can fill the gap that’s potentially left by swings away from Huawei providing equipment really quite quickly. So there’s just an enormous amount going on in that space around 5G, edge and telecoms networks. So I thought that was interesting to call out.
Jeremy Cowan 4:46
Really interesting and obviously it’s a trend that’s been brewing with Open RAN. Would you say that this is going to hit a problem with the response from the vendor community from the hardware vendor community?
Jim Morrish 5:04
I don’t think so. A number of vendors are engaging in this. The likes of Nokia, they are engaging in this type of in this approach to deploying networks. So splitting the hardware away from the software. So they are participating. Frankly, I think they have to, and, you know, potentially, maybe a sensible approach for a telecom equipment vendor is actually to split the organisation into a hardware piece and a software piece because they can have really quite different dynamics. And at some point, those P&Ls are going to have to be separated. So, really interesting times…
Jeremy Cowan 5:42
How soon do you think we’ll see that split between hardware and software, which certainly makes logical sense?
Jim Morrish 5:49
In terms of published results? I think it’ll take a while. If I was CFO of an established telecoms equipment vendor, I would have done that already. Because you really need to figure out where your cross-subsidies are, and prepare the business for the future.
Jeremy Cowan 6:06
Thanks, Jim, that’s a really interesting canter around a massive subject and one, I feel sure, we’ll come back to. Now Dheeraj, which technology story’s caught your attention?
Dheeraj Remella 6:17
It was one by Verizon. It’s how they are actually investing heavily in MEC, and the multi axis edge computing. And in my opinion, I think like this is just the beginning of the centralised role that MEC is going to play in enterprise 5G. The rationale behind that comment is that, you know, as you move intelligence closer to the edge, there is more that you can actually do at the edge. Now you can actually move parts of your core functionality, and parts of your BSS (business support system) closer to the event source, or closer to where the action is, so to speak. Right. So it could be on-prem, or network edge, or wherever be it that’s, I think the gentleman that’s leading the initiative is Adam Koeppe, (Verizon’s SVP of technology, strategy, architecture and planning).
Jeremy Cowan 7:09
That’s really interesting. So where do you think this is going to take the industry?
Dheeraj Remella 7:16
So the cost process that I’m going through right now is, you know, when you look at the digital transformation, and driven by bringing intelligence into digital twins, right, so when you look at a digital twin, it’s historically been, let’s record the state, or the state change of a physical asset. And then let our business analysts and data analysts look at what transpired, and what is this information supposed to mean to us? And what intelligence and insights can we derive out of that?
I think that has actually come to fruition at this point that we will need to start looking at how do we put this intelligence and insight back into operations and start driving the physical assets, right. So the physical assets, could be machines, people, processes, whatever be it, the digital representation is going to start driving, that the actions that are considered the next best action based on what is transpiring in real life. And one of the key things that 5G brings to the picture is the lowest latency, lowering the latency of communication, and hopefully, the service that sits behind this communication. Because this ensures that the decision and the action are not out of sync, because the world has actually moved on.
Jeremy Cowan 8:39
It’s interesting what you say about actions. If IoT is anything to go by sometimes changing the tech, both hardware and software is the relatively easy part. It’s changing the processes and the business performance. Do you think that’s going to be a harder thing to do?
Dheeraj Remella 8:56
I actually coined a word for that internally. My folks at Volt they love me for coming up with new terminology, because now I have to explain it right. (Laughter.) Everyone has heard about digital transformation. Here is the new thing. Digital transliteration. Let’s take status quo prophecies and just digitise them, right. So you see that a lot is happening in the digital transformation, quote, unquote, it’s not truly a transformation. You’re just having whatever you have today and digitising them, and hoping that you’ll all of a sudden, like start seeing operational benefits out of it. Other than laying off people or restructuring the organisation, I don’t see much benefit in that, right. So because these processes are designed based on assumptions and constraints and efficacies that are outdated at this point. You have better technology, better connectivity, better processes. You need to bring that into context and be able to rethink the original problem you were solving and rethink the answer to that, and reflect that in new and updated business processes, then digitise that, right? So, doing that significant amount of work that needs to go in to truly leverage what we have on our hands, rather than just doing the same old, same old with just a newer beast, so to speak.
Jeremy Cowan 10:18
Yeah, well, I’m gonna have to go away and think about a lot of the articles that I’ve written on transformation and whether I should have been talking about transliteration. (Laughter.)
Dheeraj, thank you.
Something that intrigued me was the report last week on the growing role of virtual reality in business. A few companies, namely people like PTC have been working on this quite successfully in the past, and the stories are on the website, www.TheEE.ai . That’s The Evolving Enterprise, to give it its full title. And according to a recent study by PwC, training employees with VR (or virtual reality) is significantly faster and cheaper than old school classroom methods. The article says workers are said to be 340%, more confident in their new skills. Please don’t ask me how they measure confidence with such precision. I’d really need an analyst for that. And where the heck am I going to get an analyst at this short notice?
Jim Morrish 11:19
I’m not sure I can help, Jeremy, on that one. (Laughter.)
Jeremy Cowan 11:23
I don’t blame you body-swerving that one. Anyway, as virtual reality in the enterprise does get more flexible, and scalable and indeed affordable. TheEE report says it’s now available to a wide range of businesses, and with workforces now more scattered, not least of all by COVID. I think any technology that delivers effective remote learning is going to be more valuable than ever. So what, you may be asking, do firms need to consider if they’re thinking about bringing VR into their own training mix. And apparently, there are three things to start with.
The first is platform selection and content creation. It may seem obvious, but your training needs to evolve with your business. And you can’t know what changes lie ahead. So don’t get locked into one provider. There’s echoes there of the TIP conversation we’ve just had. Also, hardware choices are crucial. There are different headsets and equipment for different purposes. So to support and scale VR training, the article says you’re going to have to ensure your VR platform works with as many different enterprise hardware options as possible. And finally, as so often, data is king; you’re going to need to capture huge amounts of data from training sessions. And as Justin Parry of Immerse, who wrote the article, he says think carefully about the data, you want to capture, how you’ll process it, and how to act on data insights. So if you want to learn more about this VR trend, go to TheEE.ai and search for the word “virtual”.
Now Dheeraj, we seem to be at a key moment for intelligence at the network edge. I really wanted to understand a bit more about your business, but in particular, where in your opinion, is the best place to bring edge intelligence?
Dheeraj Remella 13:42
Ah, Jeremy, that’s a really loaded question! (Laughter.) It comes down to what is the problem that you’re trying to solve? And what is the context that would make it efficient and actions relevant? You know, it could be an industrial automation scenario, or a windmill farm scenario. And in either case, you have to look at what is the problem I’m actually trying to solve, and pick a location that is closest that would help you make the best decision, right. So, you can see a variety of edges that are being discussed on-device edge, on-gateway edge, and then you have on-prem network edge. So most typically, if you want to take a broader context, then an extremely narrow version of a single device, I would go between either on-prem edge or network edge provides you in a geographic placement for having a broad enough context but to the purview of the problem being solved.
Jim Morrish 14:44
It sounds interesting. Just digging into it a little more and taking a different angle on the same topics; what advantages do you see data-driven decisioning bringing to the edge?
Dheeraj Remella 14:56
We’ve touched upon this earlier, Jim, that the digital twins are going to become a central factor in digital transformation, right. So when you look at Digital twins, these are what used to be just storing the state are now going to become the intelligent drivers behind the physical assets. For all intents and purposes, I actually consider myself as a physical asset behind the FANGs (Facebook, Amazon, Netflix, Google) digital representation of me. Anything that Google asked me to search for, I just searched. Anything Amazon asked me to buy I just buy. The intelligence is sitting behind, away from the physical asset in a digital form. And the digital twin concept is going to evolve to a place where everything that you have learned in the past, I mean, machine learning exercises, are going to sit in here within event-driven, decision-making platforms. Something that I came up with, you know, events are coming in either an individual event or a near-past context of a complex event is going to be representative of something that you have learned in the past. And by observing and recognising that set of events now you can actually act on it. And this action needs to happen in a very, very low latency manner. That 5G provides this low latency communication is going to help us arrive in bi-directional communication between the physical assets and the digital twin. That’s where I think that maximum value is going to be derived from data.
Jeremy Cowan 16:34
So Dheeraj, how do you see edge intelligence playing a role in this digital transformation?
Dheeraj Remella 16:41
We talked about the process change and that’s a really important thing. And then bringing that updated process that now has taken into consideration the new set of technologies and capabilities available for an enterprise and the new set of constraints, right. So if we are going to bring the intelligence out of individual devices, and move the intelligence away from the cloud, and closer to the event source, this digital transformation requires a combination of capabilities to play together in tandem, without human intervention, right. So when you actually take a lot of events data streaming from multiple sensors and telemetry probes, you need to be able to collect all of that data and stick it into a unified platform, given the small footprint available near the edge for infrastructure, and be able to drive complicated and complex decisions that are powered by what you have learned in your machine learning. So that’s the second technology, the event stream machine learning and understanding in the current context, and does bringing in a certain level of data-driven AI, I guess – I want to be very cautious of how I use AI, because it’s so broadly used.
I recently saw, I think one of the gentlemen that I actually follow Rob Tiffany on LinkedIn said, what is known as AI, if you remove the covers is just a bunch of cryptoanalysis. And it’s a really interesting comment, right? So when you take everything else, it’s coded by a human, that’s not artificial intelligence. But if everything else is generated by what you have learned and what you’re observing, I will put that into data driven AI, I guess. And all of these things; AI, ML (machine learning), event streaming data, and low latency complex decisions on all of the complex set of data needs to play together. And with this available, business processes need to be redesigned to take advantage of these.
Jeremy Cowan 18:47
Just one last thing that occurs to me from what you’ve said, given the amount of data and the amount of intelligence required at the edge, is this going to demand a whole new generation of infrastructure at the edge? Or can we work with what we have very effectively?
Dheeraj Remella 19:07
I don’t want to deviate from the path that we’re going because I think the right path is to software enable everything and not just build more and more specialist appliances. So, for example, like volt – I don’t want to plug my technology here, you know, but … – we are used in specific scenarios and we’re completely software-enabled and hardware-agnostic. There are several other technologies like that. And that’s the direction that we need to go into and just avoid this specialised hardware and appliance creation for specific applications. I’m not talking about the FPGAs, I’m talking about full blown appliances.
Jeremy Cowan 19:47
Thanks, Dheeraj. That helps. And where can our listeners find out more about VoltDB’s work on this?
Dheeraj Remella 19:56
It’s straight forward: voltdb.com. And there’s a lot of information that will help clarify on this new vision that the world will embark on. And we’re here to help.
Jeremy Cowan 20:10
Well, you’ve given us a glimpse. Thank you very much for that.
We’ve reached the final section of the podcast and in What The Tech! we share something tech-based that amused or even amazed us. Jim, tell us what struck you.
Jim Morrish 20:29
Thanks, Jeremy. So I found an article, which is around Amazon’s Ring drone camera. What this is is it’s released by Amazon in early 2021. It’s the Ring Always Home cam. It’s a flying camera drone released by their Home Security Division. It sits in a dock. And, if it is prompted, it undocks itself, it flies to a given room and it potentially records footage of maybe a break-in. So that took my attention. And really I had three thoughts on that. The first is just the incredible speed of development in this space. We’re looking at, there’s drones, there’s artificial intelligence, you’ve got autonomous vehicles, and these areas are just developing really quickly.
So the concept of this camera, which undocks itself and flies around your house and learns the routes around your house using AI, you know, five years ago, that would have been witchcraft, maybe even three years ago. It’s an incredible pace of development. And the next thing is the privacy issues that it raises again. And it’s just another thing sort of chipping into that, well I would say privacy debate but it’s not really being debated. It’s just something that nobody seems to have got their arms around. We’re still living in the land of multi-page terms and conditions, which nobody reads for many services. So there’s a privacy dimension too. And lastly, it fits in that sort of ignoble category of IoT projects, where you kind of wonder what’s the point? (Laughter.) It’s kind of interesting, but why? It has five minutes of battery life. I guess people would buy anything, but maybe some things should never be built. What is the use of filming a burglar when you’re not at home? Is that really what you’d want to do? Or are you just offering him some nice piece of technology to put in his bag and make off with? (Laughter.)
Dheeraj Remella 22:26
It just reminds me of this movie that I watched, where five minutes is all it takes to launch a laser-fitted shark at the burglar. (Laughter.)
Jim Morrish 22:37
Maybe. There’ll be a fortified version of it coming soon, I suppose.
Jeremy Cowan 22:42
Armed and dangerous. Dheeraj, What tickled you this week in the news?
Dheeraj Remella 22:48
Last week, I saw that 5G-connected tattooing. It’s a really good development. But then I watched the longer version of it, where they were setting up the tattoo artist. And then I actually connected how the tattoo artist is set up to do the remote tattooing. And then I looked at some of the comments that they’re like, “Oh my god, this is going to be just revolutionising remote surgery”. So you need a cadaver on the other end for the surgeon to work on. And, you know, we need to have a plenty supply of these cadavers, that are weight-proportionate, height-proportionate, gender-specific. And I was just like, okay, this is an interesting step, but let’s not go to surgery yet. I think that’s a little bit more sensitive area. I’m not sure what the thought process behind that was? Did you guys get a chance to watch it?
Jeremy Cowan 23:47
I haven’t seen it yet, but I find that absolutely fascinating. And it brings to mind the possibility of a whole new generation of what we had in the UK two centuries ago, Burke and Hare raiding graveyards for bodies that can be used for this kind of experimentation. It doesn’t bear thinking about. (Laughter.) Let’s hope we can avoid that coming around again.
Jim Morrish 24:09
I watched the short version of it. The image that sticks in my mind is partway through the development cycle when they’re connecting the remote tattooing device to the artist, and just calibrating it. There is a point where it approaches the surface of the fruit, and then plunges all the way through it. And they go, “Oh, we didn’t mean to do that.” So things really can go wrong. (Laughter.)
Dheeraj Remella 24:37
That’s exactly why I thought surgery should be considered, you know, the last part of the development. (Laughter.)
Jeremy Cowan 24:43
Yeah, I think we’ll leave that till later. Well, there seems to be a video theme emerging here because, according to the website TheIntercept.com the story that caught my eye was the FBI, which has been warned that internet of things-connected doorbell cameras like Amazon’s Ring or Honeywell Home are giving criminals early warning of police raids. I don’t know if either of you’ve seen this. The Intercept says a leaked FBI bulletin is warning law enforcement agencies of an unintended flipside to this home security tool. And that is that the cameras are alerting residents when police arrive to carry out arrests or searches.
The FBI has issued a report on what it calls “opportunities and challenges” for police from network security systems like Ring. And the FBI report was revealed in the recent Blue Leaks hack into the websites of various law enforcement agencies. Now, obviously, part of these cameras’ appeal is their ability to detect movement at your front door. The system sends phone alerts of suspicious activity. I guess this applies whether you’re in and minding your own business or just scrambling out of the bathroom window to escape the police. Because sometimes it’s the police themselves who arrive unannounced. And what puzzles me is, I’m really not sure why it’s come as a surprise to the FBI and to others, that crims might be using IoT technology. But clearly it has. Anyway, the FBI’s bulletin quoted an incident in New Orleans in 2017, and there a subject could see and hear everything that the FBI did as they moved around his house. And he was using his home security to covertly monitor the very people who were there supposed to be monitoring him. So they were using an unnamed make of video doorbell, and I’m sure nobody wants that publicity. But for impartiality, I should say, other makes of video doorbell are available to the criminal community. I don’t know what you make of that, guys?
Dheeraj Remella 26:56
Very interesting use of technology. You know, it’s like what used for monitoring. First of all, I don’t know why I need to look at the Amazon driver dropping the parcel at my door, because they already take a picture of it and send it as to where they have left it. In case the driver breaks into the house, we already have the Amazon drone that can fly out in five minutes and be able to record that, right?
Jim Morrish 27:27
Yes, what struck me about this particular story is the resultant focus that is being placed on video door cameras by law enforcement agencies. And the reality is any halfway competent criminal who actually wanted to monitor his door or his premises or anything would have just got a webcam or some other form of connected camera. So limiting the response to saying we must watch how people are using their Ring cameras and establish registers, etc. kind of misses the point.
Jeremy Cowan 28:01
Yeah, not for the first time in our experience. Anyway, gentlemen, thank you so much. Time is up. It always seems to come around too fast. Let me just finish by saying a big Thank You, first to VoltDB’s Dheeraj Remella for all your expertise, Dheeraj, and also for sponsoring the podcast. Thank you.
Dheeraj Remella 28:22
Absolutely. It was my pleasure again, Jeremy. Thank you so much for having me and letting me voice some of the thought lines that I have stuck in my head. It’s been a pleasure speaking with you.
Jeremy Cowan 28:32
It’s been great to have you here. And my thanks too, to one of the industry’s leading analysts, Jim Morrish of Transforma insights. Thanks, Jim.
Jim Morrish 28:41
Great. Well, thank you, Jeremy, thank you for the invitation again. It was an enjoyable chat. And I look forward to listening to this one when it’s published and the future versions as well.
Jeremy Cowan 28:51
We’ll get you back soon, I’m sure, both of you. We’ve enjoyed having you here. Finally, thank you, ladies and gentlemen, for joining us around the world. Don’t forget to subscribe wherever you found this pod. And be a total star, give us a 5 star rating and say something to make us blush.
Until the next time. Keep safe. Keep checking IoT-Now.com, VanillaPlus.com and TheEE.ai for tech news and interviews. And join us again soon for another Tech Trends podcast looking at enterprise digital transformation. Bye for now.