This episode of Over the Edge features an interview between Matt Trifiro and Dario Sabella, ETSI MEC Chairman. Dario is involved in standardizing tech ecosystems and coordinating internal alignment on edge computing across SDOs and industry groups. In this episode, Matt and Dario discuss the importance of standardization for Multi-access Edge Computing, emphasize why collaboration within the telecommunications industry is critical, and what’s necessary for the future of edge computing.
This episode of Over the Edge features an interview between Matt Trifiro and Dario Sabella, ETSI MEC Chairman.
Dario drives new technologies and edge cloud innovation for new communication systems. At Intel, he helps create world-changing technology to drive business and society forward. Dario’s involved in standardizing tech ecosystems and coordinating internal alignment on edge computing across SDOs and industry groups. As a chairman of the ETSI MEC, he also helps oversee the standardization of an open environment allowing for efficient and seamless integration of applications across Multi-access Edge Computing platforms.
In this episode, Matt and Dario discuss the importance of standardization as it pertains to MEC and the work of ETSI. Dario provides insight on different practices for deploying processes and workload at the edge. He emphasizes why collaboration within the telecommunications industry is critical to best serve customers, and what’s necessary for the future of edge computing.
---------
“The benefits of the coupling software and virtualizing everything, and doing so in this environment with virtual machines, containers, that's the power of this kind of cloud transformation, where MEC is a part of the journey. So, why not exploiting this flexibility, and then why not consider many kinds of deepness of the edge that are perfectly fine in principle? Of course, there are trade offs, but these are deployment choices, which depend, for example, on convenience, requirements, but also costs.”
“The standard is there for a reason, for long term interoperability and stability. But the need and approach to standardizing is not to over-engineer everything because otherwise we are risking to close the market. This is going really against also your interest.”
"To build end-to-end this kind of service, we have to collaborate, and everybody has to talk the same language in terms of data exchange practically. That's the role of the standard. So, restful APIs, data types, definitions, all the messages, everything should be in the standard. Otherwise, it's a mess. Well, in the competition, otherwise, there is no cake for anybody. Right? You talk about a total addressable market and then just having a portion of the cake of the market. If we compete and don't create a standard, there is no cake for anybody There is no market. So, the standard is enabling this entire cake and allowing them to work together because it is not actually competition, it's a collaboration."
---------
(02:50) Dario’s Path into Technology
(4:10) Who sets the wireless communications standards?
(5:42) ETSI’s role in the standards world
(07:00) Why a customer-oriented approach is critical
(08:09) Importance of industry specification groups in MEC
(11:20) Significance of a standard’s body
(13:20) Defining multi-access edge computing
(16:05) Where is the edge?
(20:55) How hyperscalers and MEC standards work together at the edge
(24:35) Exploring the MEC Federation
(27:23) Cloud and edge in everyday life
(30:45) The importance of standards
(33:20) The future of MEC (Multi-access Edge Computing)
(38:50) Dario’s goals for the future of the industry and standardization
--------
Over the Edge is brought to you by Dell Technologies to unlock the potential of your infrastructure with edge solutions. From hardware and software to data and operations, across your entire multi-cloud environment, we’re here to help you simplify your edge so you can generate more value. Learn more by visiting DellTechnologies.com/SimplifyYourEdge for more information or click on the link in the show notes.
--------
Connect with Dario on LinkedIn
[00:00:00] Narrator 1: Hello and welcome to Over the Edge.
This episode features an interview between Matt Trifiro and Dario Sabella, the European Telecommunications Standards Institute, MEC Chairman. Dario drives new technologies and edge cloud innovation for new communication systems. He helps standardize tech ecosystems and he coordinates alignment on edge computing across SDOs and industry groups.
As a chairman of the ETSI MEC, he also helps oversee the standardization of an open environment allowing for efficient and seamless integration of applications across Multi-access Edge Computing platforms.
In this episode, Matt and Dario discuss the importance of standardization as it pertains to MEC and the work of ETSI. Dario provides insight on different practices for deploying processes and workload at the edge. He emphasizes why collaboration within the telecommunications industry is critical to best serve customers, and what’s necessary for the future of edge computing.
Before we get into it, here’s a brief word from our sponsors.
[00:01:15] Narrator 2: Over the Edge is brought to you by Dell Technologies to unlock the potential of your infrastructure with edge solutions. From hardware and software to data and operations, across your entire multi-cloud environment, we’re here to help you simplify your edge so you can generate more value. Learn more by visiting Dell.com for more information or click on the link in the show notes.
[00:01:36] Matt Trifiro: Two years ago when I started the Over the Edge podcast, it was all about edge computing. That's all anybody could talk about. But since then I've realized the edge is part of a much larger red.
Pollution. That's why I'm pretty proud to be one of the founding leaders of a non-profit organization called the Open Grid Alliance for oga. The OGA is all about incorporating the best of edge technologies across the entire spectrum of connectivity. From the centralized data center to the end use devices, the open grid will span the globe and it will prove performance and economics of new services like private, 5G and smart retail.
If you want to be part of the open grid movement, I suggest you start@opengridalliance.org where you can download the original open grid manifesto and learn about the organization's recent projects and activities, including the launch of its first innovation zone in Las Vegas, Nevada.
[00:02:20] Narrator 1: And now, please enjoy this interview between Matt Trifiro and Dario Sabella, Senior Manager of Standards and Research at Intel Corporation.
[00:02:30] Matt Trifiro: Say, Hey Dario, it's great to have you on Over the Edge podcast. How are you doing today? Hey,
[00:02:35] Dario Sabella: thank you. It's a pleasure for me to be here. I love to have this conversation with you today, and thanks for inviting. .
[00:02:42] Matt Trifiro: Yeah, absolutely. And you know, one of the things I like to do is, is share some of the background of my guests, of my audience.
And I'm wondering like, how'd you get into technology? Do you remember, was there a moment, something formative that happened?
[00:02:53] Dario Sabella: Oh, well this is a nice question. It was a long, long time ago, more than 20 years ago in this domain. Now of course, I'm here having the pleasure to talk with you, representing et. The standard body on edge computing, but of course I started when I was younger as assistant specialist and as an engineer and a system on ship, small startup, and then moving to the operator domain, working for telecom, [00:03:20] major telecommunication operator in Italy, and then moving recently, five, six years ago in Intel.
And our senior manager in, uh, next generation in standard division. So it's , it's a long journey and I did a lot of changes in my career, but I, I'm happy also to, to discuss with you on those edge computing topic setting. Also the background from different other domains that I explored in the past. So thanks for the question.
[00:03:48] Matt Trifiro: Yeah, you bet. One of the things that that I think is confusing to people that are first coming into telecom, and obviously you've been in telecom for a good portion of your career, are all the different standards bodies that contribute to this global phenomenon we have, which is wireless communications.
Can you. Outline sort of the top sort of standards bodies, and then fit Etsy, which is a standard body into that, that picture, and explain how all these groups, at least the top three or four, interact with each other. So, so
[00:04:17] Dario Sabella: indeed the world of standardization is a very, very articulated and complex There are.
Many barrier in groups which are covering different areas. And the role of Etsy is quite wide. Uh, of course here today I'm just talking with you about mec, Mac on edge computing. But of course, the, the work from Etsy started a lot of years ago working on many I C T technologies starting from two G, the advent of the SIM card, if you remember, But now also covering many, many areas from the artificial intelligence, cybersecurity.
Of course, there are many other standard bodies, and of course in this entire world, it's also good [00:05:00] and desirable that also, not only within ets but also with other organizations, that there should be, of course, a good ization. This is of course, Topic, which is nowadays very important for global applicability of standards.
It's important to have a standard specification in place for every technology. It's also equally important for the market to not be confused, not having multiple standards conflicting on the same topic. This will be bad because this will have a maybe more market fragmentation. Instead, our monetization among the standards is very important.
How? How would
[00:05:35] Matt Trifiro: you. The role that Etsy plays in the standard world compared to say like the gsma.
[00:05:40] Dario Sabella: GSMA is not standard body as in just the trade association. So I will not see this overlapping in terms of work because actually we have a good collaboration with gsma. Gsma recently joined. As well, the membership of ET Me.
We are collaborating with this kind of industrial associations. This is critical and vital for standard bodies. In Etsy, Mac, for example, we are working with them perspective, for example, populated by many stakeholders, not only operators, but I would say mainly representing the boys of service providers.
And this is very important for the standard to capture the require. What is in the need from these kind of stakeholders and then put in the standards. That's our role. It's customer based approach. Customer oriented approach. Yeah. I like to talk about customers because actually they are who are requesting.
Urgently or not urgently, a standard, and then we have to do the work, which is [00:06:40] useful for them.
[00:06:41] Matt Trifiro: When you say customer, do you mean Yeah, the, the telecom providers? The, the communication service providers, or do you mean like the enterprises that might consume services that are delivered over a wireless
[00:06:52] Dario Sabella: network?
Oh, thank you for the question because you are hitting a very important point in my opinion. If you look at the membership in Etac, this is a very heterogeneous set of stakeholders. The membership is continuously growing. I don't want to do advertisement. I was very happy and proud to see in June, 2021 we were like, uh, under than 10.
Now we are hundred than 25. The membership is growing and say continuously, not just populated by a single category of stakeholders. So coming to your question, customers, who is the, the ben? Of the standard. Huh? Why are we doing this? Not just because we love papers and we love writing something useless that nobody reread.
This is another story if you like writing something, but you have to, I dunno, do something else. No, we want really to do. Only what is needed by the industry. This is my approach, not because I'm lazy , but because I don't want to over-engineer, over standardize everything. And then having a customer oriented approach is critical.
Then your question, who is our customer? If we identify other businesses, uh, if you identify your customer, then you know what you have to do. And our membership is a indication of what is the heterogeneous ecosystem that is interested to work and make. Of course, there are maybe people who are just there watching and passively attending or understanding they want to learn.
They are also. [00:08:20] Engaged companies and also more active companies. Our customers are, let's say, in two main categories. Even if of course you, if you look at the membership, there are operators and telecom providers. Telecommunication companies in a wider sense are also IT companies. Not traditionally in the domain of it.
I Cantel, for
[00:08:41] Matt Trifiro: example, is of,
[00:08:44] Dario Sabella: you mention this as a chairman, I should be, let's say, impartial. Oh, fair
[00:08:47] Matt Trifiro: enough, fair
[00:08:48] Dario Sabella: enough, fair enough. Of course, it's not an advertisement for my company, but of course Intel is a stake there. And also let's say all the companies in the entire valley chain, from service providers to cheap set producers, but also terminal side.
And why not? Smaller companies, startups. This is a key point, made me not well captured by everybody. The ISG me that I'm representing is isg in terminology, in ET terminology, ISG means industry specification groups. That means that you don't need to be necessarily ET member to join m. Okay. It's a cheaper way because you don't have to pay the annual fee, for example, if you want to be make participant, of course being Mac member gives you more rights.
I don't have the details, but if people are interested, this is a mechanism that it's took in place also to attract smaller entities and companies and startups, even everybody, but including also research centers. Governments, there are governments also in Mac. Every kind of stakeholders can be interested in that.
At every level, the ISG Mac is lowering the barrier to enter and joining this [00:10:00] kind of group, because we really want to welcome all the stakeholders in the value chain, in the ecosystem. Our customer in two categories. I can group. Even if the companies are very white, also including universities I said, and governments and smaller companies, bigger giants and so on.
But if you really want to understand who will benefit from the specification, then we have two categories. One is infrastructure owners, whoever is working on building the edge and the other side application developers. Who is going to consume this infrastructure to design applications? Those are the two bigger categories to oversimplify, of course.
Yeah. Uh, simplification is not always perfect. I always say to my colleagues in the group, we have to focus to both categories. This is a chicken egg problem. Eh, very simple to remember. The, the edge computing success depends on engaging the two categories. If you instead insisting engaging one of the two, you will fail.
Example, I start asking operators or infrastructure owners, please deploy Mac. Uh, yeah, the answer could be, oh, yes, I would love to, but I don't see the ecosystem of applications. Okay, so you, I, I go knocking at the door of application developers and they are fully busy in designing apps for the mobile.
Please, can you design something for dh? I would love to, but I don't have infrastructure, so, so at the end of the day, for the success of Edge Computing, we have to approach both kind of categories of stakeholders, and the point is what we can do as a standard. Right. So the [00:11:40] specifications, to give you an example, are not enough.
The application developer is very busy in the design of the application. They don't have the time to read the pdf, even if this is publicly available for free , and we realize this years ago, that's why, for example, we published the Open API representation of all the APIs in the ET Forge portal and Yam.
Everything else can be also not only human readable, but also machine readable. Also, for testing purposes, for automation of testing, we organize a plaque tests event for the interoperability and make hackathons, for example. Now we are organizing that we been October the final price giving day for Make Hackathon, where we invite application developer with a technical challenge to deploy an application and design application using Mac APIs.
That's more practical work. I would say not less or more important than specifications and standards because maybe the standards are instead read by. Who is going to build the infrastructure. They want to do things properly, they have to do it to comply with the standard, but that's not enough. We play this of double game because also we established it since some years or so, the code working group exactly on that direction to attract the ecosystem of developers with this kind of other complimentary activities rather than traditional.
Production of specifications. Right. Not only paper, but also other actions which are equally important.
[00:13:11] Matt Trifiro: Yeah. And that, that's probably unique in Etsy's life cycle. Before we go on, you and I are talking about Etsy. Me cuz we know what it is. But you know, [00:13:20] a lot of our listeners might not. I've sort of followed the history of Etsy, Meg, even before, like when Nokia was doing liquid apps and things like that.
And originally it was called mobile Edge computing and it was that way for a long time and then at some point it became multi access edge computing. So can you help us understand what you mean by multi access and what drove the sort of rebranding of the group?
[00:13:43] Dario Sabella: That's one of my favorite questions.
Matter ? Yeah, indeed. I believe this was a very smart move. We have to acknowledge that at the very. The MAC initiative started at the end of 2014. That was very early time. Really genuinely needed and driven by operators. It's a fact. And
[00:14:05] Matt Trifiro: telecom, wireless operators.
[00:14:07] Dario Sabella: Yes, exactly. That's why Thero mec, mobile edge computing, having in mind always a cellular network in terms of location where I can put my edge servers.
But then the industry recognized that's not. Let's not limit to the seller network. Yes, that will be a big driver. And what is the importance of cellular network, which is still there? It's mainstream. Okay, but that's not enough. Let's not limit Mac to seller networks. It's used widely by everybody, right?
And why not widening the scope and including also other kind of access technologies. So wifi and fixes access. That's what we are doing since MAC phase two in I believe, 2017. And we started not only by developing APIs, for example, [00:15:00] tailored to the practical information that you can get from the radio access network, which is still a core business of Mac, but also including APIs for the wireless land and fixed.
right? So in principle, yes, MAC is access agnostic. I can tell you for sure. This is something that was requested by the industry. I'm very happy of this move. I believe it was a wise choice, a wise move, and also by keeping connecting to what I said previously by keeping. Seller networking. 5G is a mainstream because it's obvious that we are seeing that.
Of course, the advent of MAC is also coming together with a stronger, let's say, emphasis now together with the arrival of 5g. But if you look, uh, definition of Mac, it's a really interesting, there is a sentence in the website. It's a multi access edge computing. MEC offers application developers and content providers.
Cloud computing capabilities and an IT service environment at the edge of the network. What kind of service environment? IT service. IT Service. Environ got it. It, okay. I'm not telling you at the edge of the mobile network, at the edge of the network, full stop. In general, and also this cent is quite smart and competent because it telling you, when
[00:16:18] Matt Trifiro: you say edge of the network, do you mean the access network and not the on premises?
No. Can
[00:16:24] Dario Sabella: be. Can be, can be. Okay. That, that's interesting because if you look at the specification, guess wifi could be Yeah. Yeah. And by the way, we are not telling you I'm ing where the edge should be. Exactly. Even in the, in the cellular network. I have to be fully [00:16:40] honest also by absolute implementing, making the cloud.
But it's possible in principle, any kind. Level of deepness of the edge is possible because at the end it's cloud computing, but that you can put wherever you want. I'm telling you why it's crazy because there are two advantages of the Mac. One is the location, right? It's a real estate problem, right? . So giving you the close proximity to the user.
You have benefits in terms of latency, bandwidth, and so on and so forth. But the other benefit. Collecting measures and input and data and information close. There at the edge. I was gonna ask you about that. That's why it would be absurd to have the edge somewhere else. Of course, it's a kind of trade off.
You have to decide where you want to deploy Mac. But in principle, if you look at the standard, there is no specific requirement for a certain level of the edge. This was a choice since the beginning from operators, Mac phase one. I want to have this flexibility since me is based on. I want there to have this opportunity to adaptively decide also, because of the workload and the needs and the different, I dunno, requirements in terms of latency.
I want to have the degree of freedom to choose where I want to put my workload at the edge. How far should be? This is, uh, intentionally not told in the standard. I know that can be let's a disappointing answer, but when people is asking me Where's the edge, often there is a foggy answer. Like, this can be wherever you want.
[00:18:16] Matt Trifiro: So . Yeah, yeah. No, no. I've, I've, even though my [00:18:20] podcast is called Over the Edge, part, part of my shtick is, you know, it's, it's, you have to be very clear about what you're talking.
[00:18:27] Dario Sabella: Yeah, yeah, yeah. So this is maybe spreading confusion, but honestly we have to give also the honest answer. It can be a really opportunistic location where you want.
to deploy edge. That's also the benefit of decoupling other software and, and virtualizing everything and doing this in an environment and built machines, containers. That's the power of cloud transformation where Mac is a part of this journey. So why not exploit in this flexibility and. Why not consider many kind of deepness of the edge that are perfectly fine in principle?
Of course there are trade offs, but these are deployment choices, which depends, for example, on convenience requirements, but also costs because it's obvious. It's kind of a very straightforward, you can imagine when if you move. To the edge. You may have benefits in terms of reducing the latency end to end because you're closing to the end, end user.
But of course you have a problem in terms of scalability. Bigger number of servers need to be maybe deployed to be there. Maybe also possibly co-located with the ride base stations, for example. And then this is coming with a higher costs. So there is a trade off course. This is up to infrastructure owners.
Having a look to the, what is the traffic. Maybe also over time during the years, they can change their deployment, the flavors and options over time, depending on the situation and the convenience. We don't want to mandate a certain [00:20:00] specific deployment option unless this is, uh, needed. Critically from a standard perspective.
So in general, our approaching standard to be open as much as we can just to open the market, not impose any kind of specific deployments, because also you can recognize that that can be different. Also, conflicting needs or preference from different operators or stakeholders. And now also they are collaborating with upper scaler, so there are many possibilities.
In terms of make deployments, you
[00:20:27] Matt Trifiro: mentioned that that sort of the vision of Etsy is to create a application developer friendly Yeah. Cloud environment out at the edge. I'm simple oversimplifying with that. I would say that if you asked an executive at Amazon or Microsoft or Google what they want to do with their cloud, they would say something very similar.
So can you help our listeners understand how those two, two statements can be true, or maybe they're not. Maybe they're in conflict or maybe they're in harmony. How does all that like actually work? Like, not conceptually, but how would a hyperscaler and. The me standard work together out at the edge.
[00:21:03] Dario Sabella: Let's not hide the, the more, most interesting questions also for the people here listening at this podcast.
So let me also add another ingredient to the game. Also open source, because it's typically the case that I know very well. My colleagues and standard delegates, which are very proud of what we do in the standard specifications. They, they like think, oh, there is. Some of them not, of course everybody, but if you are, let's say, very skilled in producing specifications and standard, you can do it optimally and make me think, okay, this will be like a holy Bible that everybody will [00:21:40] use.
This is not the case. That is an entire world outside standard that doesn't care about standards. Let's be honest on that. Starting from open source implementation. And Proprie implementations and other stakeholders that are maybe not, still not adopting a standard. The standard is there for a reason for long term interoperability and stability.
The need and approach to standardizing is not to over-engineer everything. Because otherwise we are risking to close the market. This is going really against also your interest. So what we are doing, for example, with, let's see, you mentioned that some stakeholders like aws, I had also some chart and left edge event where I was in a panel with those kind of other stakeholders.
They're actually very interested in edge computing and we are working with many operators, which are also doing trials on Mac. In collaboration with uppers, I can tell you also we are working hard with GSMA and O ppg. Group O PPG stands for operator platform group, and this is populated by many stakeholders.
There is a project also in a CNCF called Camara. On open source related to a big concept called it the Telco Edge Cloud. In the standard, we call it Mac Federation. It's mostly the same thing. Basically why? I'm telling you, because this is a joint interest from. Many global operators and also hyperscalers and partners on edge computing.
You just
[00:23:10] Matt Trifiro: released a white paper on that, right? Yeah. On the federation, exactly. In
[00:23:13] Dario Sabella: June. Yeah, I saw that. Yeah. This is just a white paper, not the standard. The standard will come soon, but we started one year ago by [00:23:20] publishing a first. The report on this Exploring the ME Federation, this was triggered by many operators.
It was dreamed by K D D I supported by telephonic about from other operators and many other technology providers, and this was requested strongly as a study by gsma. We started then this journey in collaboration, not only with gsma but also with the three gpp. We are working together to put standards in place to.
Provide an answer to their question. Please create some standards that will allow interoperability for this TUR Cloud concept. And we call often also in the standard bodies about federation. It's not an ownership of anything. A federation, when there are multiple operators, they are just collaborating and joining forces and sharing infrastructure and edge resources to each.
And also with the partners like hyperscalers, whatever is the infrastructure. There can be in principle, very diverse ecosystem of partners, very ous. Some operator will come along, some other with an hyperscaler, with different edge infrastructure, and they can decide what they want. To share in the federation.
It's like a party, huh? I bring my Coca-Cola, you bring, uh, your other drinks and we go together with the same, the same party and then,
[00:24:40] Matt Trifiro: and, and federation for a second. Yeah. So, so it's, it's a word that some people are very familiar with. Sure. Let me try to describe how I hear it, which is you have different operators with different business models, different funding, different countries, maybe different locations around the planet, different resources, and you're saying, You can present a [00:25:00] common pool of resources potentially across multiple operators that a developer could consume.
So I could write an application once and have it run. Globally, potentially, if enough people adopt the standard. Is that what you mean by federation?
[00:25:14] Dario Sabella: Please remember, Matt, that this is not just a concept that I invented myself or I Mac. If you look at many years ago there was, let's say a cloud computing concept, let's say defined by an I S D and also the Cloud Federation concept.
What is I D N I S D. Oh, I, okay, got it. Institute. Yeah. It's a American institution of a standard technology, so it's the clouded. Is a defined concept by N I S T and based on that you can understand if you want to move to the edge, you can understand what can be the Edge Federation or the Mac Federation.
[00:25:51] Matt Trifiro: Now is Cloud Federation the same thing as multi-cloud kind
[00:25:55] Dario Sabella: of. I always see Cloud Federation multi-cloud. Maybe yes. But now there are also other terms, like the operators talk about telco h cloud or operator platform terminology is . Very various and so we, we can get easily into confusion, but the point is really, yes, the opportunity is.
As an operator, A or B, I can exploit the infrastructure of someone else in a certain location, and having my app hosted by the infrastructure of someone else to be closer to my customers is like a sort of Mac roaming. If you are an operator, you don't have a certain footprint in account. Your customers there in holiday, I [00:26:40] dunno, from Spain to Italy.
They want to enjoy, they have a sim card from Spain, but they are in roaming and there is also the opportunity to give a sort of service continuity. Okay. That gives them the feeling that, of course they are just connected to another age, but they don't care. This is just a technicality for the end customer.
The essence is, They can have this kind of service continuity and enjoy the usage of a certain endpoint application for their terminal that can be closed to their location, whatever they are, and the Federation is a tool for that. Because can be by direction. There are business agreements, of course, among those kind of stakeholders.
It's up to them to decide what they want to share in which conditions and so on. But there is a mutual benefit from doing that, joining forces and sharing infrastructure resources with the benefit for the end customer. This is a great market. An opportunity for everybody. Of course, this is coming with a big complexity.
I can tell you for sure.
[00:27:40] Matt Trifiro: roaming is, is sort of genius, right? I remember when I had my first cellular phone roaming didn't really work or you'd end up at someplace and it cost you a lot of money. Roaming today is just brilliant. Like it just works. It makes a lot of sense to me that I, my phone is used, or I'm just saying my phone, it could be any device, but this is, that people can understand.
It could be my car, it could be my, my computerized license, whatever. Oh, your car as well? Yes. Yeah. But, but the fact that I can consume the same applic, My device will operate the same way no matter where they are in the world, is very compelling, but it's also super compelling for the developer, cuz that means my addressable market is gigantic.
Like if, if I have to write a [00:28:20] separate application for every single operator, you know, certainly I got one for Vodafone and one for Verizon and one for telecom atal, like I I just that to make my head explode. But if I can run one. App and have it run across a federated met cloud. That that's pretty exciting.
And I do think it's super complex, but roaming was super complex and so if anybody's capable of solving it, it's the telecom.
[00:28:41] Dario Sabella: And you're totally right. Matt, you mentioned also the car. This is one key example where this kind of use case is very critical. Talking about,
[00:28:49] Matt Trifiro: can you walk us through like a really, really clear use case that, that like, that normal people could understand?
Yeah,
[00:28:56] Dario Sabella: sure. There are multiple tires, so with multiple operators you can find in many price also from, for example, 5g a. I know this because I'm working as well in five ga, but this is just the public information. But automotive also because 5g, it's in the membership of Mac and they have a urgent need or put some standards in place for that.
So one of the examples and use cases, which are also by the way, captured by GSMA in their O PPG group, is on edge resource. Know the sharing. And typically if you, I believe this is very straightforward to understand also for everybody when a typical example of in a smart city and the cars, you cannot imagine in the city there will be only one single kind of cars.
Everybody with a. BMW or with a Ferrari with a, it's not realistic, right? We have to think about, first of all, there are multiple car OEMs in this scenario, and you cannot pretend to have everybody, all the cars connected with a single operator. This is not realistic again, so in real life you have multiple.[00:30:00]
Operators, multiple cars, maybe the sync card connect, connected with the multiple operator network. And then of course, each of them can have their own MAC vendor and network infrastructure vendors. So there is a heterogeneous scenario three uh, dimensions, multi m and o, multicar oem, a multi vendor. This is the level of complexity, three dimension level complexity.
This kind of heterogeneous system. There needs to
[00:30:25] Matt Trifiro: be a standard for that
[00:30:26] Dario Sabella: ex . So can you join Mac please? Because you know exactly what, what is the. So they, they are in desperate need of a standard for that because they are asking, please, please put in standard desk kind of things because for longer term benefits, in terms of interoperability, I believe this is the key word, system of stakeholders.
To build end to end this kind of service, we have to collaborate and everybody has to talk the same language in terms of data exchange, practically, that's the role of the standard. So restful APIs, data types, definitions, all the messages, everything should be in the standard. Otherwise, it's a mess well in the competition.
Otherwise, there is no cake for anybody. Right? You talk about a total addressable market and then just having a portion of the, of the cake of the market if we compete and don't create a standard. There is no cake for anybody. There is no market. So the standard is enabling this entire ca cake and allowing them to work together because it is not actually competition, it's not collaboration.
That's the approach of the ET mac with this kind of [00:31:40] heterogeneous set of members and participants, really, not just because we love collaborating, but this is also need from the industry. They need that.
[00:31:49] Matt Trifiro: I might wanna drive my car from Italy to France. Right. And I want, I want, I want all of the capabilities, whether they're safety oriented or navigation or entertainment.
I want all those to go with me. Yeah, exactly. That makes a lot of sense. So I wanted to come back to something you mentioned earlier. That's actually a. Pretty passionate interest of mine. I call it observability. I think you may have called it something else, but I work with a lot of developers that build applications at at the edge that are very, very latency sentence of very, very jitter sensitive, very, very timing sensitive.
And one of the things that they need that isn't obvious to most people is they need to be able to react in real time in autonomous ways to conditions of the network. At the millisecond level. Right? And so one of the challenges, not just wireless, wireless was sort of a black box for a long time, or, you know, certainly all that observability is controlled by the, by the operator.
But also in, in data centers and wired networks and all of that, like that has been something that, that whoever owned that would control it and wouldn't want it to go out. Or if, if they did want it to go out, they don't have the tools to expose it and the, the way that a developer needs to consume it.
And I know that getting access to that data, Is pretty foundational to me. So can you talk a little bit about how Mech is looking at that and, and what you see evolving
[00:33:12] Dario Sabella: there? Yeah. Thank you. This is a very interesting question and also if I'm allowed to say, I'm not [00:33:20] exaggerating, but you are talking real about a critical need to understand Well, not all, for all use cases, to
[00:33:27] Matt Trifiro: be fully honest.
No, but, but for a lot of the really important.
[00:33:30] Dario Sabella: Yeah, we have to be also very honest and, and that that's important for, for the credibility. We cannot claim for all the use cases. There is the same requirement in terms of latency, but there are critical use cases which are in need. Well, it's
[00:33:43] Matt Trifiro: not just latency, it's resilience.
[00:33:45] Dario Sabella: Of course, of course. But talking about, for example, latest, you mentioned information from the network. This is very critical and sometimes, for example, there is an api, for example, in Mac Radio network information api, the MAC 12 specification. This is start since the beginning. Of course, this is tailored to the radio access network for the, yeah.
Is there a fixed network? Yeah, the same. The same thing. There is a spec on WLAN and spec fixed access APIs. They are, they say the same, the brothers, the equivalent for the other kind of access technologies. I cannot exclude. In the future there will be other versions. And now also we are having a study item exploring abstracted information.
For application developers. So we are kind of building a developer friendly environment and API that could obstruct those information because as a application developer, you don't have to be expert on 5g. Actually, they need us to have information in time about what's happening there in terms of network conditions.
I can tell you even more. You may want to know it even in advance. Before, that's, no, I'm not joking. Matt . That like predicting the future. Oh yeah, yeah. Oh, absolutely.
[00:34:55] Matt Trifiro: Absolutely. No, no. I totally understand. But is
[00:34:57] Dario Sabella: a reality. I'm not joking. This is also [00:35:00] part of the activities in Three GPP as well. It's called in Advancer QS prediction notifications, and we have also predicted QS activity in 5G in the Association for the Automotive.
So it's a huge work you can ever look to the white papers published by 5g, aa, and now in the Mac V tweaks api. Because we have an API for video. And we are about to publisher next releases soon. Very interestingly, but already in the publisher version, there is a predicted QS data type. That's cool. The application should be able to carry from a sort of function, which is called the prediction function, which can be everywhere in the network in the term, I don't care.
Maybe done by other third parties that can elaborate sort of big data prediction analytics. Yeah, and W D A F is, that's
[00:35:48] Matt Trifiro: genius, is to build it into the api. That's really, really, that's the right way to do it. That's,
[00:35:52] Dario Sabella: that's really neat. Yeah. Can be also driven by ai, some engine that could elaborate the data series from the past and can do some predictions of course.
We cannot predict with the same reliability, different deepness of the future, but depending on the need, you may have an information about the reliability of the prediction itself and so on and so forth. Just to give you the example, for example, for the automotive side, you may want to understand if your automated and connected car feature can be activated with a certain level of acceptable reliability for the car maker.
because I don't want to risk the life of my customers, of the driver and offer a service to do that. How reliable is my wireless network? Of course, people will tell you a hundred percent being a [00:36:40] wireless. Mean, there is no situation where you can guarantee a hundred percent that the network is having a certain level of minimum quality of service.
There is always a level of confidence, which is acceptable, but in any case, you need an assessment. So for example, there are activities. In standard values and industrial association, QS stands for quality of service and you want to predict quality of service of the network. And these can provide some data that can be gathered at application site through MAC api, for example, for the B two and the application.
Knowing it in advance, this situation can. Like, for example, for I streaming, you mentioned infotainment. Uh, there is a stream, maybe there is a bandwidth needed and you'd know that you are just entering a tunnel in the mountains and, and then you just predict there will be a, a drop of the signal. But maybe you have also some buffering at application level somewhere that can compensate.
But if you know in advance you can just tune how much bandit or change. The level of the resolution for a little bit of time. Yes. Right. And in practical case, if the application has this kind of information in advance, they can exploit it and improve the quality of service perceived by the customer.
Maybe me as a customer in the car. I don't understand what's happening there. I simply see a better service and I'm, I'm happier. Right? What's happening there? It's, uh, kind of magic .
[00:38:11] Matt Trifiro: This is one of the, the sort of little, little understood and not yet very well marketed. Um, aspects of [00:38:20] building applications that run reliably at the edge and deliver a, a, an acceptable, you know, service level.
And I'm glad that, that et Cmec is, is pursuing that because it, it's like the whole thing falls apart without that. Like, it's just, it's such a critical piece of it and it's really neat that you've built in the, the prediction. So you're looking at the world through the associations that you're part of, and indirectly through Etsy, maybe indirectly GSM A and the 5G Automotive Association, all this.
Outside that world is a lot of other stakeholders, a lot of complexity. If you could nudge one aspect of this, because the entire vision that you have is slowed down by what's slowing it down, what would you try to push through?
[00:39:02] Dario Sabella: I can tell you one of our priorities also is to compliment our activities and collaborate with open source communi.
[00:39:11] Matt Trifiro: Okay, like Kubernetes, like the cncf,
[00:39:13] Dario Sabella: we are having collaboration with the LF Edge. For example, this ME Hackathon is jointly organized between Etsy and LF Edge. Oh, cool. Aino Arina is a project which is collaborating with us. Of course, being aware that this is a standard body, we are not doing something.
That is done by open source communities. Our work in Etsy, me is a standard body, is complimentary with open source, but should be really genuinely aware of what's happening. And when, when you design something, this has to be, let's say, aligned with what, what is the trend, right? So when there was the advent of the nfv, we clarified in Mac phase two already how you can deploy with the virtual machines and then the, with the advent of containers, we [00:40:00] clarified alternative virtualization technologies that can be adopted by Mac.
This was really driven by. The external world outside there and we are collaborating with open source. Of course. I would love to increase this level of collaboration because I believe this is very important. Also, et there are some open source groups which are starting. The old example was, uh, osm, open source, mano.
For the management and orchestration framework driven by the, the, the advent of an FEV and so on. But in general, I believe the standard, not only nets in general, the standard should be very aware what's happening also outside there, especially in the open source, sometimes practical implementation, working code.
Is starting from there and not, uh, from the standard actually will say that it'll be even better. This should be actually the rule. Always the standard should be driven by implementations. That will be the golden rule, not vice versa. Let's not invent something that then it's just paper. Use it by nobody.
[00:41:03] Matt Trifiro: That's what the open source like source code. Efforts have really been good about, you know, the, the phrase, the phrase in developers code wins, right? Like you could debate whether the hci api, channel eight parameters or nine, the guy that does the poll request, that's, that's who wins. Cuz that's the implementation.
I think that's very,
[00:41:19] Dario Sabella: very smart. Yeah, sure, sure. We have a collaboration with life on that since years and I would love also to strengthen this kind of collaboration. I can hide, hide my my wish that soon. Maybe they will also join, even the Mac directly contribute. Why? Right. Yeah. But of course it's a different environment, so maybe they, they have to make this code stable and then [00:41:40] also to converge in the industry among different stakeholders.
Yeah. Once they feel happy. This also part of the collaboration we are having with O PPG in Multipart collaboration, O G, SMA and Kamara in cncf, and at the gpp. So GSMA is asking us to put standards in place for that, and equally they are collaborating with Kamara and CNCF for the open source reference implementations and so on.
What is Kamara? Camara project is a project in cncf working on the open source area to cover this kind of a telco, edge cloud environment. Oh, interesting.
[00:42:13] Matt Trifiro: Is that an implementation of the api, an open source
[00:42:15] Dario Sabella: implementation? It's a, it's hard to explain one single sentence, but actually starting from the Telco Edge cloud, they are doing some early work, more urgent from open source society.
We as a standard bodies, Targeting, putting in standardizations the key reference point and interfaces for this kind of operator platform architecture, which is requested by gsma. So at the end of the day, I will hope there will be, let's say, a sort of input from open source to standards at least we are providing in our open area.
The MAC 40 specification for the Edge East Westbound interface for the ME Federation. And we are welcoming also direct contributions to the standards. So there is a possibility for that of course, as all the standard work. This is
a
[00:43:06] Dario Sabella: in contribution driven. Yeah, so it depends on contributors, but. As a customer oriented approach.
As I said, I always there with the door open for them, and I would love to see inputs on that because that's really a [00:43:20] genuine sense of feeling that there is an interest and a need to put in the standard something. So we are just doing what is needed and requested by customers. In this case, GSMA for defining the interfaces.
Once there will be a more detailed definition of the APIs, they can join and implement it vice versa. For example, there can be, as you mentioned, adopting MAC APIs. Yes. This is something that is, for example, defined in Aino project and there are different projects and blueprints which are using Mac APIs and MAC frameworks and specifications.
There is a interactive mind map in the their API portal where they are using. Make standards and APIs, this is happening. So also vice versa is possible and this is a nice collaboration that we are having with Arina since a lot of time.
[00:44:12] Matt Trifiro: That's great. Hey Dario, thank you so much for giving us a view into this world that I think very few people understand and obviously more should, because it's gonna be very critical in us evolving these applications that we wanna deliver at the edge.
So I really appreciate you coming on the show. Can people find out more about Etsy, Mac or get involved?
[00:44:32] Dario Sabella: Sure. Of course they can contact me. , I, I can be happy to facilitate of, or they can go on the etsy.org website or type just in the search and Gene Mac, multi access edge computing. That is a website for our, our group, which is the public portal, et.org/technologies/.
Multi access edge computing. Right? And they can find everything there, all the information. That is a nice video [00:45:00] clip on talking about what's actually Mac, the specifications, the other initiatives in the area, white papers that you mentioned, hackathon information. Tile, proof of concepts, all the initiative that there are, and of course if they want to join Mecca.
I believe this is a great opportunity. I believe this is the momentum. I see also growing interest in that from many companies. And of course we will love to see inputs because actually this is the value if you want to join and we want to influence the. That's, this is the
[00:45:33] Matt Trifiro: moment. Excellent. So thank you so much, Dario, for being part of our show, and maybe we'll have you back in a year and you can give us an
[00:45:40] Dario Sabella: update.
Certainly. I would love to meet you again. Thanks for inviting Matt. It was a pleasure. Thanks so much and thank you for your attention. That does it for this episode of Over the Edge. If you're enjoying the show, please leave a rating and a review and tell a friend. Over the edge is made possible through the gen sponsorship of our partners at Dell Technologies.
Simplify your edge so you can generate more value. Learn more by visiting dell.com.