 Good morning. Good afternoon. Good evening. My name is Medasa Ahmed. I'm principal with Mitra Corporation. My main interest is 5G cloud and cyber security. I've been involved with the Linux Foundation for a couple of years now. I joined onapp.com looking at the CI CD security for software assurance and also looking at how we can use onapp for security automation. That kind of introduced me to a couple other projects. I started working with Elephantac and got engaged with 5G Super Blueprint. So today's talk is really how can we accelerate 5G innovation? How Super Blueprint can help the community, the ecosystem? To get ahead in an evolving marketplace. My background prior to coming to Mitra was truly in networking and cloud applications. I worked with a major vendor in networking technologies for 18 years and some of the folks have been mentioning how they contributed code to open source and sometimes they pulled out. I've gone through software development, software testing. I also at some point in my life managed an incubation lab and that was one of the interests that brought me to Super Blueprint. The work was very similar. An incubation, a bunch of engineers we used to just think, can we do that? And would it solve that problem? And we had the budget so we will put that team into a lab, let them figure it out Would it work? Would it bring any efficiency or optimization? And can that be productized? Into some sort of either product or service? So I thought the Super Blueprint has that signature. That's why I joined Super Blueprint. Right now the way I see the networking and 5G self is changing the market. Market is in sort of a flux, right? New changes are coming. Some of the companies may be already on board and they're prepared to tackle those changes. Some are not. So 5G itself is a complex system and I put this slide together to share you know, what is the complexity? And if you look at the network evolution, we went from 3G, that's when the data was introduced. LTE provide a lot more data capabilities and 5G is all about data, right? If you look at the components of 5G, the IMS system, IP, multimedia subsystems, they're still the same as LTE, right? So there's not a whole lot of changes. But the changes came in where software is decoupled from the hardware. Then the software is also distributed. They also made an attempt to separate control plane from user plane and that kind of trickle down into when you look at 3GPP standards for RAM, radio access network. Or later on when ORAN Alliance tried to pull together interfaces, data standards, message exchange. All of this brings in a lot more complexity. The good part behind those complexity is that it still requires the technologies, the Linux Foundation or any enterprise has been using, right? So in order to deploy, and that's kind of where I put this graphics in the middle, is there are a lot of common layers that we're already familiar with. The complexity that comes in is that 5G brings in new deployment architectures, new interfaces, and if you look at from an architect or designer perspective, it's just another application, running in the cloud, virtualization or hardware servers. But they all have to work together. So there needs to be connection between these layers. For 5G system to work, you also have to have these BSS and OSS applications, right? For any system to run, you need to have observability. You need to have automation. So when I look at 5G, as Arpit early on in his keynote mentioned, 5G is really a digitization effort of telecom industry. And to me, digitization is that you reduce the friction between service consumer and service provider. In technical terms, if you're an application, you're deploying a container. Your infrastructure is providing the infrastructure services to the application. And if you're deploying a Kubernetes environment, maybe your network is providing the network or IP backhaul services for your different deployment enclaves. And if you just dig down, you might come to, you know, there's some vendor who provides dark fiber optical transport for you. So there are so many layers where one layer is providing a service to a vertical layer. And this all has to work well together. Services, when you think of 5G services, it's still going to be those three common services. You still have to provide SMS. You still have to provide voice. But the major change is on the data side. The 3GPP introduced in release 15 onward a new deployment architecture for the code. And they also defined how radio access network is going to be distributed. But they did not go into the details how you will instantiate these, right? So it leaves it open. You could have container-based. You could have VM-based. And you can actually put in the old box. But all of these are possibilities for an operator. They have to use those. They may have base stations out there where minor hardware changes can take that base station to open-run software. They may have a core. And they are already the operator who are using hybrid core where, you know, the system takes the LTE function and also communicates with the 5G function. So these are a lot of possibilities out there that create a lot of complexity. On the other hand, technologies we are using are not new. So we have expertise. We just have to bring it together in a systematic way. So it's skilled and resilient. So going back to our talk of the day. What is 5G super blueprint and what it is not. We talked earlier. There are so many different softwares out there. In LFN community and LF community in open source, they provide capabilities. So if I look at super blueprint at a very high level, I like to put it into two buckets. We have industry challenges and we have network challenges. And going back to my previous slide discussion. The network challenges are how do these different layers talk to each other? If you think about communication service, if you get an order for an enterprise customer, they want their 50,000 employees to use 5G and our guest here, our host here, tell us. And they are on the hook to actually fulfill that order. You have to have some workflows and systems that take that order, push it down through their back office business application to some OSS applications. And those OSS applications have to then look at the design of the network where this user is going to be. And then provide the code resources, provide the radio access or air interface resources and then be able to route the traffic to the right destination for that. So those are the network challenges. On the industry challenges, one of the big things that the 5G promise was IoT. IoT meaning there could be smaller devices that collect data. And provide that data back to an application that can intelligently control that environment. And this graphics you could see V2X, that means vehicle-to-vehicle or vehicle-to-controller application, which we already seen in Tesla Model and a couple other EVs. They're doing some of these vehicle-to-vehicle communication, but it could be collaborative with the municipal use cases. For example, you're going on a beltway or a ring road where you want to adjust the traffic flow, you could uniform or you could get that visibility. In factory, this could be your factory robots. So quality measurements could be one of the use cases. Arpid earlier mentioned agriculture use cases, right? Be able to deploy weather sensor, humidity sensor, water sensors, et cetera, that can optimize your field management. So for any operator, you have to look at the use cases. You have to do assessment and prioritize, do the planning, then integration and orchestration. This super blueprint group of technologies or capabilities kind of shows you that their capabilities are there. We just need to put them in some order where they can provide an end utility to an operator. So today I'm going to kind of touch a few of the things why super blueprint will help accelerate the innovation. I'm going to go over some time-to-market cost-selling and operability use cases, et cetera. So time-to-market is another thing, right? Since market is developing new solutions, if you are not part of it, if it takes longer for you to deploy, I'll give you an example in maybe five, six years ago, if an operator had to increase the bandwidth, they had to buy the optical channels. They have to configure a whole bunch of routers from point A to point B. This whole work with maintenance schedules and the cost, cost could be 100,000 plus per gig for additional bandwidth, but it may also take three to four months to get it in place so you can actually use. That's just one example. You could have a whole bunch of other complexity, for example, deploying your workloads. All of those, if you can do that sooner, you may be able to offer that service to the consumer. A lot of time, building a service, advertising a service to a customer and having your capability in place may have a lag. If you can do it, you may lose that opportunity if operator advertises that I'm able to offer this service starting Monday and all of a sudden calls come in. Those customers are not going to wait, right? They may be able to find alternate solutions. 5G Super Blueprint started on bringing together some of these capabilities to offer how you can do some use cases. Originally, Virtual Switch and Virtual Firewall was a use case, but then we expanded. I think 2021 Super Blueprint team actually offered a demonstration how we can deploy a 5G core and ran network to offer 5G services. And then last November, we demonstrated an IoT application. So we're going from network challenges towards industry challenges. An industry challenge was how can we use IoT? Part of IoT is you have a massive sensors network that talks an application. We were able to show a safety inspection in an industrial zone. For example, the demonstration was about industrial zone requires safety equipment. And the example was you got to wear a hard hat. And we deployed the camera. Application was able to pick. A person coming in the zone would be able to pick whether or not a person was wearing a hard hat. So I think all in all, if you don't do things on time, you may miss the market opportunity. On the cost setting side, I took this example. I'm a bike rider. So I thought I should explain what the cogs are. So on the bike side, smaller cogs make the bike go faster. But in order to use that capability, you got to have either low incline or very high fitness level. And we in our different organization, we have a different levels of maturity on a lot of these technologies. So it's a very relevant example. You can go slower, you can go faster if you use the right gear. Cogs on the business side has a different term. It's the cost of goods sold. So in the accounting term, every projects have to have, you know, it sounds financial house straight. And the cost of goods sold could be if I want to launch, let's say, a 10 gig service, what would it take? Right? So part of it would be capital expenditure. You're going to bring in software, hardware, all the miscellaneous items that go with it. But you also have a development and integration cost. In a service provider network, you know, when you develop these services, you call it product engineering. They need to pull together all the package. Once deployed, it will work as intended. And whatever needs to be part of that deployment architecture, that has to be engineered together. Right? That's what TELUS was talking about, how they pull together things. So those are the costs that will go in. Then you have operational costs. I remember when I first job back in late 90s as a system administrator, we had about one to 12 ratio. If there were 12 servers, networks, windows, Linux, Solaris stuff, there was one engineer that will manage patching maintenance, you know, all that kind of stuff. But things have changed, right? Now you can manage thousands of servers. You just need one person. But why do you have the right type of automation and tools in place? And we're going toward machine managing the machines. You guys here in LFAN and open source community are makers. You make that automation software. So last thing I'll talk about the cogs is, you know, all the cogs don't matter. If return on investment is negative with service provider, the project is not going anywhere. So in most cases, those will be assessed. There will be so many different analysis done on the financials before even project takes off. We can do that. We can lower the risk on ROI, maybe increase the amount of ROI or decrease the time, right? Because if you think about it, if you buy something in the U.S. at least, you get five-year amortization on those hardware. You put the expense, you bought the things, and you can write it off for tax purposes for five years. After that, you can't write it off. So if your deployment takes about a year and a half, you only have three and a half years to recoup your expenses, right? So somewhere you need to accelerate those things so that the projects have a maximum benefit of return on that. So another thing that I usually give people examples, and I think that example fits well here, LEGO's example, right? So LEGO's are different blocks. When we were kid, we'll pull together different colors, build something out of it. I see that open source produces a lot of these LEGO capabilities, but they need to come together to become something. Whether you want to build Star Wars, that's the one common set that you can buy with LEGO's. Build a city out of it or create a Tonka truck, whatever your favorite is. But in our case, I think we need to pull together these different capabilities LEGO blocks so we can offer a value-added service to our end user. In those cases, there will be enterprise customers. There could be service provider customers, right? In LFN, we have a whole bunch of projects. If you take the Umbrella Linux Foundation itself, it covers cloud, it covers networking, it covers a lot of endpoint utilities. So diversity is important because if you are not dipping into open source, you may be facing intellectual incest, right? You have a very close environment, limited number of options. Your own company, your own center of excellence is limited to those ideas. When bringing it to LFN projects like Super Blueprint, you get a different perspective. And those perspectives may bring an option that's more optimized, less expensive. Same thing with scale. You may be able, your product or your project may be limited currently. When you bring it to Super Blueprint, that project may have wider integration scale. And you can also discover interoperability. For service provider, these things are important, right? They always want best of the breed. In the past, if you've noticed, for every capability, you will have minimum two to three vendors. Because they need to make sure that business survives if a vendor files bankruptcy or liquidate. So they cannot go with the best of the breed. They will always have to do three. I think that's the same thing. If I'm trying to solve a problem with them, let's say Super Blueprint, we earlier talked a little bit on nephew and own app and couple other automation and orchestration tools. I think they can all coexist because they may provide that hedge against if project goes into archive load, right? Another part of interoperability is the 5G is not new, right? Most people, except for the enterprise, which will do a greenfield deployment like large enterprises who will be putting their own networks. But for major operator, this is not new, right? It has to interoperate with their existing system, their LTE system, their VSS system, OSS system. So interoperability is another way to look at does your product, your project, or your capability interoperates. So a little bit of background on the use cases. We, early on for 2023 planning, we looked into what can we do this year under Super Blueprint. A couple of philosophical decisions there, does it have to be open source? We know the reality that for a service provider or a user, there's a mix, right? They will have vendor solution, they will have open source solution. So we are open, at least in the community, that both open source and vendor solution have to coexist. So based on that, we're looking at we got a lot of capabilities within Linux Foundation. We want to bring them in to solve some of the network challenges. Once we build a baseline 5G system, then we can use industry challenges and provide solutions for industry challenges like we did for IoT last November. So we invited a couple of new companies, the Joint Linux Foundation, to see what kind of use cases we can come up with. So one of the use cases that we have right now is workload placement on mobile edge compute. And that's a real need. The 3GPP standards allow you to distribute certain radio access network like CUDU and the RIC, right? So there's a non-real-time RIC and real-time RIC. Real-time RIC has some latency requirement because it provides a lot of optimization, machine learning and correction and mitigation for the radio access elements like CUDU and RU. So we think that that kind of need will demand that there is some way to figure out how to deploy certain workloads in certain locations. How can we do that? So the first use case, MAC placement, we are pairing with Equinix. It's a cloud data center edge compute provider. They have some tools. We're going to bring those tools with 5G deployment tools that we have in our super blueprint. Another use case that we are considering is remote attestation for IoT devices for security and authentication. So a little bit of context on IoT device compared to PCs, smartphones or servers. We expect IoT devices will have very skinny operating system. IoT devices in some cases may be just a small device that sends beacon and beacon is coded for binary bed decisions. Is there a moisture or not? So it will send a very small message that needs to go back to the IoT application. So it may not have a lot of capability like a full-fledged operating system that can offer where you can put something. So this particular use case, we have a new open source project from BatterTone Labs. They wrote a software called sediment. So sediment is a validator verifier engine and a small code that goes on IoT device. So when IoT device boots up, it's going to send its passport or its credentials. And if the credentials are verified, that IoT device becomes part of the IoT application. So that will give you security on when you're deploying mass number of sensors, you're deploying cameras or any other IoT devices within a particular industrial use case. Reason why it's important, because as we increase IP endpoints, these all IP endpoints can be jumping board for somebody who has nefarious intents. And these IoT devices will not be able to use access lists and things like that, right? So it's important to know who you're collecting data from, who's part of your network. Second part is IoT devices onboarding amendments. So if IoT devices can provide you a point in time information about their security and their credentials, what happens when you change? Because if your config changes, your software level changes, that may change your own persona or your own capability. So those have to be accounted for. And finally, we're looking at changing the super blueprint to provide network as a service. And as we've seen, there are multiple layers of 5G as an application, provides communication services for data transport, voice and SMS. But then it consumes resources underneath. So we're taking that approach, network as a service. There's two parts to it. One for sure, we're going to do network as a service and try to create demonstration and provide some recommendation on that. The other part is, can this be a self-service portal? Can these services offer a portal or API where you can connect these layers together? And I think ONAP has some capability because you can do design, orchestration and deployment. Nephew has some capabilities, but you have to offer some sort of playbook. Super blueprint does not write codes, right? It is an integration effort to show this can be done, then document that. So if somebody else has to do it, they don't have to spend days and days trying to figure out how do I get the network up or that particular domain up. So our intent is that every program increment cycle will offer a service. That service will include documentation, URL to those libraries or repositories. So if somebody else has to do it, they can do it in short of time because we figure out how to do it with all the mistakes that were made. They've been taken care and now you have a correct procedure. I think that will axle right because there are a lot of people who wants to do different services based on 5G network. But if you go by a 5G network from Alex and Nokia, any of those, it's a huge expense. So think of this, if these playbooks or these software packages can provide you a lab in the box. The people who are researching new capabilities, they can easily do it. Or even if you are writing some optimization code, you could do that easily because the rest of the environment is available for you to experiment with. And that will come in handy because part of that, these use cases, we have done a slicing demonstration previously. But what we expect that in the future, your slice needs to carry your security SLAs. And that way you can deploy secure communication, resilient communication as part of the same workflow that you're using for your instantiation of particular network function. So this one, I think all of these points have been previously discussed. I think between success and failure is going to be can you take your project or capability and introduce it to the wider audience? And can you provide usability in terms of how to make it happen? I think a lot of these projects, like we're looking at ODL, not to put you on point Robert, we have to take the capability and present it in a way where it's easily usable. And that part we can gain through here where different projects can come in, whether you are doing a networking as a function, where you're virtualizing a particular communication layer, or you're adding automation for observability or mitigation loops. All of these, you know, a lot of these capabilities are there. We just need to show how this can be done so people can start using it. Because bringing, you know, when I looked at that slide that Arpe put today, there are so many projects, right? Where do I start? That information is now there. With the super blueprint, our intent is that for use cases, if you want to do industrial IoT, here's the playbook. You take all these software, deploy it, and then you tweak, right? It's a learning process. I don't think anything is perfect right now. But between success and failure, if you don't have learning, we can't cross that bridge. So we need to get these things out, and that's beneficial for different projects. They already have their capabilities, just not touching the, maybe, right issues or challenges, by bringing those into super blueprint to demonstrate how a particular challenge can be solved and solutions can be provided. I think those learning can improve the success rate and accelerate the innovation. In the past year, year or two, we had a couple of companies like Klum. Klum is an all-round compliant UPF and firewalling company. They're brought in, and we integrated that into our solution. GenExcom GXC software was used as a distributed base station for 5G antennas. So we used their CUNDU in our solution. IBM provided IoT application servers. That's what we used to demonstrate that you could do industry inspection. You know, once you have that blueprint, you could use it for anything. I think we can repeat that for agriculture or any other municipal. For example, if you have a few sensors out in the city, in the wintertime you want to know the snow or humidity. I think you can get that information easily. The Paraton is bringing us the attestation. Right now we're focused on IoT because that part hasn't been touched. Servers and PCs have endpoint protection. So there's some way to attest server and PC, but IoT devices hasn't been touched. So that was the focus for the group. We're using that in a coming demonstration. VaveLab is helping R9 network help orchestrate and deploy 5G in 2021, and they're still active with us. Equinox is providing mech support how to deploy workloads on edge compute. And Equinox is another company that open source a software called Cube Armor. Cube Armor is a security software for Kubernetes systems that can expand to other. It provides you visibility to communication between functions on the network. So it's kind of based on EVPF, but they created policy server, policy implementation point. So if you build a slice in 5G, you deploy the network. You can actually deploy the security policy, which network function can talk to each other on what board, and it will monitor and control. So I think what we're looking for is key use cases for secure deployment, for slicing, edge compute, service creation, deployment automation. And we heard today, a lot of these capabilities are already an elephant portfolio. So this is where I try to reach out to projects and the companies who are in this business either as a user, because we could share ideas, identify needs, and then if you have resources, you can bring your software for different projects, or a company who wants to integrate with 5G Super Blueprint, they can bring their software, they can bring their expertise, or just bring your leadership, right, how we can get this next use case, next increment set up. I have some information here, Super Blueprint at LFNetworking.org. And then there is a URL underneath on the slides that will be shared with you. You can click on it. We have a huge backlog, because we've been collecting the use cases that we intend to undertake in the future. Your support will be very helpful. How are we looking at the time? All right. Any questions? My question was, would you maintain the Blueprint even after releases? Let's say that you released a couple of Blueprints earlier with Arnaud, you, Amcop, and you orchestrated an edge-to-cloud connectivity in one of the Blueprints. So as time goes on, every component has migrated to the higher level versions. If somebody wants to have released all these Blueprints, if I want to go and try that Blueprint, now the components are already at a higher level. If I start doing the Blueprint, I have to either go to the lower level because they won't support those versions anymore. So you will try to make sure that all the Blueprints are up to date? Yes. So Blueprint is slightly different from software release. So we don't produce code. I understand. When you say Blueprint, my understanding of Blueprint is you are trying to connect all these different softwares together to solve a new case. We are showing that proof-of-concept. So let's say from here from now I want to try that proof-of-concept in my lab. As you said, I want to try it out. But all the software you talked about has already moved a few versions up. They all bumped up. So I cannot go back to the older Blueprint and then go try it out because the dependencies are not met. If I try to pull the software for that, it's all changed, the API changes. It might not work as specified by the Blueprint documentation. So there's a remote possibility that might happen. But because we're looking at the capabilities, not a strict API. For example, we use IBM, right? Next time we don't have to use IBM if there's another software available. We can use it. What you get from Super Blueprint demonstration and outcome is set of instruction, a write-up, a white paper. We do demonstration, right? It's a POC demonstration. This is how it works. We capture in that white paper is, what did we use? Sort of software bill of material, hardware bill of material. What are the key configurations, right? When we use GXC with open 5G core, we capture the configuration on GXC. How a CU is configured to connect to a free 5GAMF, which is to us, which is important fact, right? So tomorrow you can use UE ran sim. You can still use the same configuration. So the way we're trying to do is to get you going by reading the white paper or looking at the outcome of our demonstration to get you started. So you don't spend two months, three months bringing up the environment. I definitely agree. That is going to be still helpful. There's no doubt about it, right? Try to bring up something. I was just worried about you talking about the configuration and then it goes to upper level. Then it's all changes. That's where the problem is. I was trying to understand whether when you have a blueprint, whether every release, maybe you have a small cycle to up the version numbers and then publish any changes to it, right? Yeah, so that will happen too. Because we do upgrade to the latest, greatest when a new version comes up. But in this sort of mix, we have... So the key components we have is code and ran, right? That makes up the 5G. Then somebody comes along like Klum came along and said, we want to offer you UPF. So instead of using Free5GC UPF, you take that element out, don't configure it, use Klum UPF. That's one thing that 5G offer is you can plug and play any of those functions. Like your UDM can come from Oracle, right? Or it could be an open source. Or your AMF, SMF and UPF could come from Magma, which was another project donated by Facebook, right? So we have the ability to do this plug and play. So we're not actually following the software versioning aspect. We're just giving you the playbook that if you want to get started, here's the software bill of material. This is the kind of hardware we use. This is the configuration we use. That will get you to a working solution. And then you sort of learn and provide feedback. Maybe next iteration we can use your ideas, right? Yeah, I wanted to point out that Super Blueprint doesn't have a downloadable and a runnable deliverable. It's mostly a documentation project which brings up the ideas. And when you say a playbook, it's not an Ansible playbook. It's a playbook for humans to interpret. And essentially tailor to their requirements, which brings it together. That having been said, the participating projects, which are actually software projects, those bits and pieces that get used in Super Blueprint end up being software deliverables for those projects. So they tend to have CSIT readily deployable containers that you can just download. It's just that we are not going to give you an Ansible playbook that if you point it to Kubernetes and I'm making things up, it will bring up a fully functional 5G core for you. No, you have to do those integrations still yourself or hire an integrator who will do that for you. I apologize. In American football, all coaches have a playbook. That's how they want to run the game if all goes well. Any other questions? Well, I appreciate your time and your attention. And I hope to see your voices in our calls, bring capabilities, and you don't have to contribute a long term, right? That's why I mentioned that software expertise and leadership sometimes is just an idea. The biggest part is finding the problems that exist that most people can say, yep, that's a true problem. We don't want to spend our resources on a problem that nobody cares about, right? So we need to sort of prioritize, find a problem that exists out there. We thought edge compute deployment is a big thing. We don't see a lot of solutions out there. We can clearly be open source and be able to provide the same utility. So any other challenges, you can shoot an email that, hey, what about trying to do this use case? One to two lines describing the problem, and you don't have to spend a lot of resources. And if you're more interested, you know, other skills can be used too, right? You can bring your programming skills to automate some deployments. That's fair game. Well, thank you very much.