What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
From Edge to Core: Reimagining Enterprise Networks in the Age of AI
Interested in being a guest? Email us at admin@evankirstel.com
The networking landscape is experiencing a seismic shift as artificial intelligence transforms traditional connectivity into intelligent, business-driven infrastructure. Gary Sidhu, Senior Vice President of Product Engineering at GTT, takes us on a fascinating journey through this evolution, explaining how networks are evolving from "dumb pipes" to strategic "business compasses."
At the heart of this transformation is GTT's Envision platform, which connects enterprises to their data and applications globally through edge and core infrastructure components. Sidhu articulates how AI is enabling intent-based networking, where administrators can simply state business goals like "prioritize video conferencing" and let intelligent systems handle the complex technical configurations. This capability allows businesses to become more responsive and agile without requiring specialized networking knowledge.
Perhaps most compelling is Sidhu's exploration of the emerging "agentic ecosystem" – a world where autonomous AI agents will generate unprecedented network demands, potentially requiring "infinite bandwidth." This future presents both enormous challenges and opportunities for enterprise networking, necessitating new approaches to traffic prioritization, edge computing, and security frameworks that can authenticate non-human network participants. GTT is preparing for this future by building its own "AI factory" with dedicated GPU infrastructure and sophisticated information architecture that powers both internal operations and customer-facing capabilities.
Ready to prepare your network for the AI revolution? Discover how GTT's approach to networking security can transform your business infrastructure from a utility into a strategic asset that provides actionable intelligence and adapts automatically to your needs.
Crossing Borders is a podcast by Neema, a cross border payments platform that...
Listen on: Apple Podcasts Spotify
More at https://linktr.ee/EvanKirstel
Hey everyone. Evan Christel, really excited for this chat today with the leading networking and security as a service provider about the impact of AI and how it's going to both transform its operations as well as its solutions that it delivers to customers. Super excited to welcome Gary Sidhu, svp Product Engineering and Technology at GTT. Gary leads product innovation and development to advance the company's solutions. Really excited about your background and expertise, gary welcome.
Speaker 2:Thank you so much, excited to be here.
Speaker 1:Excited to have you and we're going to jump into all things networking security AI. Going to jump into all things networking security AI. But before that, maybe, gary, tell us about your background and bio and your journey to GTT today?
Speaker 2:Absolutely, Evan. As you mentioned, I am Senior Vice President of Product Engineering at GTT. At GTT, our mission is to connect people and machines to their data and applications anywhere in the world, and my job is to have the technology to make that happen in GTT. At GTT, we believe the Envision platform is that technology which simplifies the complexity of managing and solving large-scale network connectivity solutions for our global enterprises, For your audiences.
Speaker 2:Envision that I'm mentioning here as three components and I'll be referring to this interview. Envision Edge it's our flexible configuration platform that can be delivered globally anywhere, what we call golden images Think of like iPhone type software already preloaded, and these images are of our network and security functions that can be virtually activated and changed on demand. And Vision Core, another set of virtualized infrastructure where we deliver our centralized SD-WAN gateways and firewall and it's connected to our Tier 1 IP network as well as hyperscale and network. All of this thing is stitched together through our AI-powered site-centric experience that lets you design, code and manage your large-scale network on Envision Edge and Core. So my role as an engineering leader at GTT is to simply empower our engineers so they can continually innovate, build and evolve Envision platform to our feedback. I've spent a lot of time at IntelliCo 25 years and a lot of the stuff that you see is all about learning and evolution, where our industry is evolving, and the topic of today, AI, is definitely very, very relevant to our industry.
Speaker 1:Let's dive in. I think you would agree it's the most exciting time in our lifetimes when it comes to AI, particularly in networking, where there's a real renaissance, a reinvention, a disruption happening that's going to, I think, reshape enterprise networking and bring along a whole new raft of capabilities. But what do you think is the most interesting opportunity that we are presented with right now?
Speaker 2:I love this question, evan. It is like you said is so relevant to the changes. First of all, I truly believe that AI-powered network fundamentally shifts the paradigm from network as a dump pipe, if I may say, to network as a business compass. Let me explain that a little bit more using an example. Network is powered with the real-time intelligence and with that you can easily identify the surge in your network activity, map it to the application causing that and you can correlate oh, maybe it's due to our new product launch and I can easily shift the resources proactively to that demand. So it's not just a pipe, it's mapping the needs of applications and the resources that you need.
Speaker 2:The network that provides actionable insights directly influences business strategy. It's not just about IT operation. We're very excited about two things with AI. One is when, using AI, you can define high-level intents or your business goals. Network automatically configure through long time industry term intent-based networking. So AI actually powers that. So instead of you as a network engineer going in and setting quality of service for each and every application, you can just declare the intent hey, prioritize all video conferencing, like the session we are having, and AI engine dynamically routes, adjust the bandwidth, meet the latency, and this makes your network more agile and your business. The result is the responsiveness of business increases, right, the result is the responsiveness of business increases, right, the one area that is fundamentally changing in a network is the evolution of the agentic ecosystem. Right, we know that the future is autonomous, goal-driven AI agent that works with applications, services and other AI agent to achieve the goal.
Speaker 2:Think about that All of a sudden, you have more than 8 billion humans on the earth generating the application traffic. It's almost to reach meeting to a point where you can have infinite bandwidth generated by these limitless agents. Your traffic is going to get exponentially increased. How do you prioritize that traffic? How do you provision that low latency compute at the edge where AI engines are running? It's a huge opportunity, and the other aspect to that is which naturally comes is in the agent tech field actually comes is in the agentic fair. We already have a framework to identify humans, to connecting authorizing applications. How do you identify the agents? So security needs to be embedded into the network where humans and machines cooperate on the network. So I think those are two big opportunities and truly in 10-based networking, you define your business goals and also the agentic web, the opportunity that it brings. It'll be fun, for sure.
Speaker 1:Fun for sure. Yes, as telecom geeks, this is fun stuff and you have an amazing network through organic growth and acquisitions. Maybe describe the network a little bit for the viewers. But also, how are you and where are you embedding AI into your network?
Speaker 2:Absolutely so. Gtd is tier one IP back home. We have 450 plus pops all around the world and we have 26 locations worldwide and it's growing. Where we have this Envision Core infrastructure around the world, it's pretty well like we have an ability. We have thousands of partners through which we can deliver the connectivity of the edge to our Envision Edge platform. Truly, you get a platform with GTT. You get a platform starting at the edge and connecting to the core and also connected to the hyperscale network.
Speaker 2:So the AI I'll speak in two angles one from an infrastructure perspective and the functionality angle. We spent a lot of time this year thinking and investing in what we called our AI factory. That is the powerhouse for our enterprise AI at GTT. Security of our data and our customers' data it's non-negotiable at GTD. So what we did was we invested in our own AI farm, so our own GPUs that we deployed in our data centers, and that gets you the infrastructure. The next level in the infrastructure is we spend a lot of time this year building the information architecture of all the data that we have the data that our network is generating, the data about our customer, how we're provisioning them, the data about all the engineering artifacts that we have at GTD and with that information architecture we are able to sort of create knowledge graphs or knowledge bases and map it to the LLM so we can find quick answers from them. And on top of that we built what we called a framework of hierarchical agents that can operate on this data. So that covers our infrastructure.
Speaker 2:What does that mean, functionality wise, for our customer? We believe that this allows our customers, our employees, our partners, to achieve faster and smarter outcomes. They have a better ability of their network performance and the generative AI embedded to those knowledge maps or graphs. You can find answers in a more non-telecom way, like a lot of our customers find answers in a more non-telecom way. Like a lot of our customers, you know networking is not their core business, but you can find that information about your performance, how they're spending at the site, how they can evolve their solutions.
Speaker 2:You can ask if I'm designed, if the network is designed. Look at my network, how it's designed, recommend me what should I improve with? So a lot of this. These are all capabilities that we continue to evolve to our Envision DX platform that will be continually built and evolved and made it available for the customer. The other part, as the automation is growing and we talked about envisioning all of these are virtualized environment. As we are adding more automation, our goal is to make these capabilities available as intent-based networking solutions so you can integrate with your AI agent all of these capabilities as well. So those are traction that we're heading toward.
Speaker 1:So much opportunity. It's such a win for customers. I love how AI is going to really enhance how you protect and defend against evolving threats so needed, as all the bad guys now have AI as well. But you're also leveraging AI to enhance customer experience, customer satisfaction. Maybe talk about that a little bit.
Speaker 2:Yeah, absolutely. Our Envision DX, as I mentioned, is all about. A lot of areas, are powered by AI already. Right, if you're a customer, like I said, you can write, ask in an easy English-like prompts that we all are familiar with chat, gpt and other AI agents. You can define that. You can find easy information about our network, your site level details about your network. And this is all powered because we invested on a component that we call a component that we call a software component. We call observability agent. What it does is it's a software that is sitting on an Envision Edge and it is streaming every few seconds the health of the machine, the health of the interfaces, and it's connected to a stuff that we already built, the streaming architecture we already built at GDT for our core network. So now you have a pretty good visibility of the packet leaving your site to the packet being delivered across our network, and all of that experience is available to Envision DX and the efforts that we are putting in here.
Speaker 1:Incredible, and it's not just the customer-facing external teams that are benefiting. You're using AI internally within GPT for lots of value-added use cases. Maybe talk about that and the work that's kind of behind the scenes to make you more efficient, productive and more.
Speaker 2:Absolutely. When we started this journey one of the things we wanted to we quickly realized that developing AI within our product should be a natural outcome. First thing, we have to build AI at the enterprise level. Right, and the more time we spent in our information architecture that I talked about and mapping them to the AI functions, it empowered our employees. First, right, we just we internally we just launched our CPQ function that allows our employees, for example, our sellers or sales engineer, to design a solution. You can type in hey, I wanted to design a network in using our Western US and they all are branches, use that template, so it can design that network, that coding for you very easily.
Speaker 2:You, we are also using this for our operations, our ei ops. Right, as this data is available to our employees and it's also available to our customers. The same amount, same data. We have a little bit more internal knowledge because of the coordinate for connectivity and other factors. So that information, the centralized information architecture, empowered our employees to find anything about the customer, which customers are coming up with renewals. Our finance can use that same use cases to give SLA credit proactively if we had some outages or any other scenarios. So I think I believe that really empowered us from helping our employees, our employees helping to our customer as well, as we also have all our internal knowledge, hr finance Also. We are continuing to map into the same architecture. So finding information to as simple as hey, we're a global company is today a holiday in the UK. You know, finding that, instead of going into some SaaS platform, find 10 clicks down and you find that AI can find you really easy information.
Speaker 1:Incredible. You have so many customers, thousands of customers, great stories, anecdotes and case studies. I don't want you to have to pick any favorite children here, but can you share a couple of use cases and how customers might be deploying AI-based solutions or otherwise?
Speaker 2:Using the GTT infrastructure.
Speaker 2:Yeah, of course, we haven't made yet the AI services available to our customers. This is to power the services they have bought. But on our roadmap we do believe the opportunity we have is putting the two things one, making the GPU infrastructure available at the edge and at the core right. That certainly we see a lot of our customers asking us for that because they have the same problem like securing, you know, running the low latency scenarios is always there, so they wanted to run their AI workload more natively on their infrastructure. So that definitely we're seeing a lot of companies have AI strategies. It's easier said than done to implement, right. So we definitely also are definitely talking about, you know, as we are able to deploy these AI workloads internally for our needs, could we make a blueprint available for our customers so that, whether they use our AI, our infrastructure or they get another, like, how could you get started with that piece?
Speaker 1:right.
Speaker 2:So making that and, I think, also thinking about how we participate in our GenTech web ecosystem. I think that definitely will help our customers. So at this point we don't have any AI services like that customers buy from us, but we're definitely thinking making that available on our infrastructure.
Speaker 1:Yeah, and all kinds of new workloads coming online every day. Given all of this experimentation and research and other kinds of trials of Agentic AI, do you have any recommendations on best practices for getting started exploring agentic AI solutions or leveraging some of the advanced security modalities that you built?
Speaker 2:No, absolutely. I think we have a framework. I definitely happy to talk a little bit about it. So, first of all, readiness matters, and that starts with data and automation. So those are. I love this quote. I love drinking wine, so good grapes good wines.
Speaker 2:Good data and good automation is such a foundational element. So I think those magic if you have a good information architecture about whatever use cases you're trying to build and you have an automation to manage that. I think that is a very foundational start. Focus on that. I think that is one. Making sure you have a well-defined use cases. It's such a shiny object. We went through our initial phases. All of this looks cool. I can find information quickly. Those are good phases to learn, but for a proper implementation you should have a proper, well-defined use cases.
Speaker 2:I think having that would be a second step. Having a flexible framework all the way from having where would you use infrastructure? How you will build Forgive me, I'm going to go a little technical here. One of the technologies MCP servers, right. Model protocol what technology would you use to build these right? And the orchestration framework the supervisor agent having a whole hierarchy of agents managing other agents, and I think that's such an important part to figure out early in the game. Technically Right.
Speaker 2:And people People are like you know, as people are transitioning to AI, the more your organization is starting to use these tools, getting hands wet, better it could be. And how you even talking, going further, talking with your HR, the responsibility of a digital worker and a human worker those are your next level, high level. You know the evolution once you're involved in that part. Making sure that sort of policies, your AI policy, how digital worker and human worker operate those are like high level framework things. Make sure your data automation exists, having a well-defined use cases it's got to be a top-down support right. Your board down, everybody needs to be bought in and your use cases have to be well-defined.
Speaker 2:And your engineering team building the frameworks all the way from information architecture how do you expose, as an MCP type, technologies, how the agentic frameworks will work? And getting that hands-on and decide your AI factory Are you going to run it on-prem? Are you gonna run some cloud services? Right? I think that would be. For us, it was very clear Data privacy was most important thing, and so we invested in our own infrastructure, right? So I think those kinds of things building your AI factory would be very critical and getting your people ready, getting your HR policies, depending on how far you want to go with AI. Getting that piece is such an important aspect.
Speaker 1:Wow, great advice, great insights and congratulations on being ahead of the curve and I can't wait to see the solutions that you roll out over the next weeks and months. Very exciting.
Speaker 2:Very excited too.
Speaker 1:Congratulations and thanks for joining. Thanks everyone for listening, watching, sharing this episode and look forward to keeping in touch. Gary, see you soon, I'm sure.
Speaker 2:Likewise. Thank you, Good luck to you and your audience.
Speaker 1:Thank you, Take care everyone. Bye-bye.