There’s always plenty written about which networking technologies will be hot as a new year kicks off. Rather than join this crowded space, I want to look at 2025 from another perspective: Instead of what will be hot, who will be hot? It’s pretty obvious that networking is going to change, but the vendors are as big a force as the technologies, maybe even bigger. From a vendor perspective, if we think of networking as the playoffs we have in American football, the competition is between three teams.
Team one, which we might call the home team, has only one player, which is Cisco. Don’t count them out, though, because they have home field advantage. They have the highest level of strategic influence on enterprise planning of all network vendors, and they’ve tuned their portfolio to fit today’s network needs better than anyone else, because of that influence. The only thing that can defeat them is change, which is what both the other teams—in different ways—are going to promote. Think of the competitors’ strategy as a new formation, a new play.
As Team One and the leader of the current market paradigm, Cisco faces a unique challenge. To look beyond network technology for change, they’d have to admit networks can’t change themselves. They’re winning the current network game, and it’s always hard for a winner to suggest changes and risk being turned into a loser. They’d need to buy a visionary company, and could they recognize one, or properly nurture it inside a network-incumbent culture? Cisco led network vendors in promoting IoT, and they have offered servers for years. But Cisco has always wanted to wait for a market to prove itself before entering—the “fast follower” strategy.
The second team has two players, Broadcom and HPE/Juniper, and their play is a draw play. Networking as an independent element of IT is done, they say; it’s now totally subordinate to IT. If you can control IT, you can trap networking in the process, so you let the incumbent giant Cisco defend its networking space while you dance away with an IT win. Broadcom has IT virtualization, virtual networks, and super-effective switching chips. HPE has strong IT influence too, and with Juniper would have a broad network portfolio.
In Team Two, we have a more direct connection to applications, and that means the players could actually drive a new IT paradigm that would ramp up network spending. The problem for them is that they’re both hyper-focused on AI, and while AI will surely require some fast data center networking, its focus is to improve information quality and not drive information volume. To expand information volume, we really need new information. Where do we get it? From the real world, meaning from IoT. Yes, it’s time to include that factor in our network game. Broadcom, because of its chip-centric product line, might be the Team Two player with the best shot at leading an IoT transformation, but will they move to exploit the opportunity in 2025 when Wall Street is rewarding them for their AI opportunity?
The third team is running a chaos play, sending everyone downfield and hoping the sheer number of choices will let you find a winning one. Everyone else in the network space is part of this team. For some, specialization is the answer; don’t try to grab too much, but instead find a niche. For others, offer a broader alternative to the other two teams, in the hope that their maneuvering against each other will weaken their hold on buyers.
In this third team, we also find vendors most interested and active in the IoT opportunity, two of which are usually not considered enterprise vendors at all: Ericsson and Nokia. Both these companies are actively pursuing new IoT strategies, and while they are surely hoping that broad public-sensor IoT will deploy and create a service provider network opportunity they can address, both are also emphasizing private 5G, which is an on-ramp to an enterprise IoT strategy. In any event, public-sensor IoT would surely drive major changes and open major new application opportunities that would drive new enterprise network needs.
It’s pretty clear that change is the real issue here. Will networking change, or will it remain as it was? Will something new come along to add to the size of the network spending pie? Nothing new? Networks commoditize completely. Team One wins by shrinking itself—layoffs. Team Two wins by exploiting commoditization. Broadcom can sell chips, but with no new traffic, not much new revenue is available. HPE is tied to a lower-margin networking business. If there’s no change on the table, Team Three is looking for commodity niches in a commodity market, not a happy situation either. In the no-change scenario, Team One wins a hollow victory.
Networks deliver information that’s created, maintained, and analyzed by applications. If anything is going to transform networks, it has to be something that tied to a different kind of delivery—a delivery of a budget. Business benefits, in other words, and a lot of them if we want a network transformation funded. We’ve been looking at network technology instead, and looking at it all wrong. You don’t invent a technology and then try to fit it to as much as possible; the candidates will all be status-quo elements that won’t generate transformational budgets. Coloring outside the lines, way outside the lines, is essential. Find a major benefit, then find what can address it. The 2025 winner will be the team that does this best, and changes the game for us all.
There is hope here for networking, and network vendors in general, but realizing the potential means stepping a bit out, or perhaps up? Above the network stack, we’ve always recognized the domain of the application. The advantage network vendors could have here is that it’s pretty likely that there is a strong network connection to the two most promising “transformational” applications, because both rely on a form of IoT, and a lot of it. And, of course, a lot of IoT means a lot of new network opportunity.
Before you think I’m trying to restart something that, in tech years, is maybe as old as dinosaurs, let me point out that the original notion of IoT, which was a broad and universal collection of sensors and intelligence, has never been realized. We’ve done the easy IoT, the stuff contained in buildings, instead. Not surprisingly, that hasn’t lived up to the lofty goals initially set for IoT. The real IoT could generate billions in network opportunity, in just two broad applications, and it’s in reach to all our teams, though some will have to reach further than others.
So who is that winner? Cisco’s incumbency and take-no-risk strategy is likely to make it the most successful player next year, and if the question was “who wins in 2025” the answer would be Cisco. But there’s that pesky “and beyond” to consider.
Cisco’s approach is a major risk beyond 2025. Remember the “change” point? There’s an enormous amount of money on the table if any player can open a new set of benefits that will drive an increase to IT and network spending. If one does, then fast-following won’t be enough, or even possible. HPE/Juniper has a shot at being that player, and so do Ericsson and Nokia, but the former has to digest major M&A in 2025, and the latter two are trying to enter a new space serving a different customer base. There’s a chance Cisco could win even in the long term, if everyone else fumbles. I think there are too many competitors for that to happen.
I’ll go out on a limb and say that Broadcom is the network winner, looking beyond 2025. Their dominance in high-performance switching chips means that a commoditizing network space wouldn’t hurt them too much—they’d fuel the commodity vendors’ gear. They’re the only company that’s seeing real AI growth from the “agent” applications, they have chips to support IoT and would benefit from growth in IoT applications, and they have, through VMware, data center credentials.
We’re entering a new era in networking, one in which the goal will be to add benefits to the business case for networks, not just deliver data that’s already been processed and factored into the budget for maintaining the status quo. It’s a new problem for networking, but far from unsolvable, and there are vendors who realize the need and can be expected to offer a solution. Think of 2025 as the transition year, the last of the old network…and the start of the new.
Wait! What are those two transformational applications I mentioned? Well, you’ll have to wait for my next column to find out.
Source:: Network World