In my role as CTO, I have the opportunity to regularly meet with interesting customers who are dealing with the digital transformation of their enterprise. They are often perplexed by the myriad of paths available to deliver better business outcomes. Early in the discussion I often ask what seems a very simple question, “Which company do you believe is leading the networking industry?” Very frequently I hear the same historic and woefully incorrect answer.
If you think about this question from the perspective of many notable historic contributions and proprietary technologies and the myriads of complex legacy protocols that are part of our day to day lives in networking – then I could agree with this knee-jerk default response to my question. If, however, you consider which company is driving the technology and changing the networking industry to a more resilient, more secure, more agile future that enables the delivery of better applications and better software to power change and improves business outcomes – then the answer is the Big Three – meaning the cloud titans Google, Azure, and AWS.
Yes, it is true that the installed base of legacy networking gear comes from a titan of yesteryear. Much like the correcting Selectric (have you ever heard of it?) was once the installed base leader for typewriters, before the advent of the word processing era, PCs, and Microsoft Word. But the titans are dominating the evolution of our networking future. The reason I ask this otherwise inauspicious question is simple – to engage in a dialogue that awakens the customer to stop thinking in the context of a CLI in the rear-view mirror of yesteryear, but to look forward and anticipate the future of networking. This is important because infrastructure investments not only have to last a long time, but customers should be on the lookout for companies and ecosystems that are in alignment with the evolutionary forces driving our industry. Just like the wrong DNA that doesn’t adapt to market forces ultimately declines or even goes extinct, so too do networking road maps and technology directions that don’t align well with the paths being blazed by the titans. Moreover, being on the wrong path will ultimately put an enterprise at a disadvantage.
Let’s take a closer look at just a few technologies and ecosystem considerations to illustrate what I mean. Let’s start with Linux. Something like 2/3rds of all workloads in the cloud run on Linux. The Linux open-source project is now ginormous and the rate of growth in its sophistication is directly attributable to the many contributors that do so on behalf of their cloud-centric workloads. Generally, the titans embrace open source communities and are often major contributors, if not founding contributors. What would networking and application deployment be without Linux?
Another consideration – containers. Though Docker did their best to take the early container capabilities of Linux and make them more consumable, Google gave the world Kubernetes; and the world has jumped on board. Containers are the basis of how Google has been running its own networks and applications for years. As the external world took notice some many years after Google, Kubernetes became “a thing”. It quickly amassed a huge community and sucked all the oxygen out of the room for other approaches to containers. It also pointed the way for everyone else to the evolution of the old school virtual machines. Containers are THE way to make hybrid cloud networking actually work. To whit – even VMware has now announced aggressive support for Kubernetes and containers.
Now you are saying to yourself, “all well and good, but that doesn’t sound like the switches and routers of networking stuff”. So, let’s talk data centers, SDN, and automation. Google was first to put a form of SDN to work and at massive scale. They didn’t use legacy networking OSs with an old school CLI or pseudo SDN proprietary solutions. They built the Andromeda SDN project. This project has subsequently influenced the direction of the world’s largest telecom carriers. At the same time, the data centers for many of these titans have been influenced by large scalable networks using the guts of the world’s largest network, which we all call the Internet. The Internet works day to day because of BGP. That is how many of the titans have built their data centers.
Automation. This is a big topic. Each of the titans has taken a slightly different path to automation. Yet, each has taken an approach of resiliency, simplicity, agility, and continuous development and continuous deployment. The vast majority of the time they are using industry-standard silicon. In today’s classic enterprise data center, aside from how VMware might cooperate and “manipulate” the network, in general, the software that underpins the network itself is very rarely updated by enterprise customers. By contrast, a titan may be making many changes per day. Granted, most of them are very small and very contained – but they are making these changes on an otherwise gargantuan network.
An example might help here. Most people routinely update their phone’s network OS. No, it isn’t a network OS – but, I don’t know about you, but my phone is incredibly important to me. If I realistically thought that a bad OS update was likely to brick my phone or knock me off the Internet or from cellular services for a day – I’d never update it. This is the situation for many enterprise customers today. However, as the enterprise network OS evolves and begins to mimic the microservices model of the titans, the full CI/CD dev process of the titans, and the levels of the automation of the titans, over time enterprises will begin to trust these new models and ultimately reap the same rewards that the titans have achieved; and that we, as their customers have derived.
Here’s the logical question: if the company you had thought of as top dog driving the networking industry is really the Selectric leader of yesteryear – then who are the actual industry players aligning with the cloud titans?
This is my vantage point:
One of the industry leaders is moving to a software subscription model but is really just repackaging old products and maintenance with new proprietary campus fabric and proprietary chip sets that lock-in their customers. Moreover, part of their product family follows one model, while the upstart part of the product line is essentially incompatible and follows a different model. Though they may occasionally leverage Kubernetes, the general technology strategy is very repackaged and yesteryear-oriented. If a company wants a subscription play, to be aligned with the cloud titan direction, it should build cloud or private cloud software that is developed and delivered in a cloud-native CI/CD model, so customers can subscribe to something that improves weekly. This, . . . sadly, is not their model.
The relative newcomer in the space that has been kicking the former leader’s tush has a very cloud ecosystem-inspired model, but their business is so highly concentrated on titan scale and F50 data centers that the trickle-down investment and effort for campus and enterprise edge networking is so insignificant that it isn’t worth commenting on at this point.
As for Extreme, we have a strategy highly-aligned with leadership technology coming from the cloud. We start at the bottom with chip sets inspired and forged from the cauldron of competition that is a hyper-scale data center. Our network operating systems have been modified to leverage the very latest Linux distributions and capabilities derived from cloud workloads. Our premium model products offer an embedded “edge computing” Linux server with a v-nic connection to the switching infrastructure. Our fabric technologies leverage open standards and aren’t built on proprietary methods. Our development model is in the final stages of transition to a full cloud-native CI/CD model.
Our management interface is cloud-centric and available from different public cloud services and can be implemented in a private or hybrid cloud. Our next-gen cloud infrastructure investments are naturally all container, serverless, Kubernetes, and service mesh-based as well. The cloud itself can directly use the leadership AI/ML tools of the hosting cloud service to provide best-in-class capabilities. Said more succinctly, Extreme is becoming the leading light in delivering cloud-driven networking from the edge to the cloud, inspired and derived from the best networking technologies of the modern leaders of our industry. Can we do even more – of course. And that is the journey we are on.
In summary, as you ponder the simple question about the true networking technology leader, recognize that tomorrow’s enterprise will be a direct recipient of the ecosystems, the tools, the opensource, the networking chips, the automation, the operating systems, the languages, and more; all of which are being driven by the cloud titans. Aligning with DNA that is consistent with the philosophies and directions of the titans, but without lock-in, puts your enterprise on the best path to benefit from the gargantuan ecosystem driving the true future of networking – the titans.