We’ve all probably read predictions about how networking will go this year, and which vendors might be expected to win or lose in the competition for enterprise budgets. How about for the years beyond, though? Most enterprises try to look ahead three years, so they tell me. How might the future they see impact the competitive landscape in the future? Is it even possible to say at this point, with technological, political, and legal factors all swirling madly around us? Well, let’s try.
I’ve received comments about trends over the next three years from 294 enterprises. All of them seem to agree that there isn’t likely to be a major change in the mission of networking for their company in that period, and so agree that managing network cost while improving its quality will be the mandate. At the same time, almost two-thirds think that their network model will “evolve” during this period, and this group believes that the evolution will likely impact their choice of vendor. In fact, more enterprises (199) think that they will likely see some changes in their vendor selection in the next three years than think their network model will evolve (195).
The big driver in vendor-shifting, according to enterprises, isn’t a change in the model but changes in the vendors themselves. Of the 199 who think they’ll change up their vendor commitments, 144 say the primary reason is to get “more stable” or “more predicable” business and product strategies, ones that align with each enterprise’s own perception of their needs. Almost all enterprises see vendor reorganization, mergers and acquisitions, or big product strategy announcements as a threat. That’s hardly a surprise; if you’re trying to sustain or improve without spending much, you’d love stable product and supplier choices.
What’s also interesting is that the 99 enterprises who do see their network model evolving are just as interested in vendor stability as those who see no such evolution. This suggests that enterprises don’t think a network model change is necessarily going to drive them to a different primary vendor. Is AI positioning by a vendor then unlikely to change the rules of competition? Sure looks like it, but let’s take a bit deeper dive.
Forget whether a model shift or a mission shift is likely or not. What is the primary driver of change to the basic model, and baseline budget, of networking? Answer: Application changes. Network planners may talk about looking ahead, but their vision is focused on what they actually can control. The thing they can’t control, and must accommodate, is a change in the applications being run, or the way they’re run. The compute side drives the network side, absolutely. This is reflected in the fact that, when I ask enterprises, the vendor they pick most often as the one driving the strategy that decides where networking is going is IBM, which doesn’t even make network products. In fact, none of the top four vendors (IBM, Microsoft, HPE, and Dell) picked as driving the strategy bus are really network vendors, and the fifth-place (a nose behind Dell) vendor (Broadcom) isn’t a traditional network product source. The top network vendor (Cisco, of course) ranks sixth.
If you link this with the AI positioning question, it’s easy to see that a presumption that a compute vendor on the top-influence list would likely have an AI position, and that network vendors may simply be trying to ride their coattails. But unless a real compute-side AI trend develops, and ends up driving network demand, AI positioning won’t change the competitive position. But maybe, in a sense, it already has, in the fact that network vendors have had to acknowledge they lack the strategic influence to control their own destiny.
One solution to this, obviously, is to merge with one of the big compute players, which Juniper hopes to do with HPE. That cedes the point of strategic influence to those vendors, but what other option does the network vendor community have? Here’s a basic truth to consider: You can’t have strategic influence if your products have no real differentiators, if you’re a commodity. The only answer is to focus on the fact that almost two-thirds of enterprises believe the network model will change. What will drive that? What specific technology and product changes will result? Answer those questions, and answer them in a way that makes business sense to buyers, and you have a shot at strategic engagement. Answer them insightfully and you have strategic influence.
This is why jumping on AI is a bad idea for network vendors, particularly the bigger ones. Whether AI is a good thing or not, it’s someone else’s thing, a topic you can’t lead and drive. What do you say to a buyer question like “Why do I need to AI-enable my network?” Do you say “Everyone know that” or “Because IBM/Microsoft/OpenAI say so?” Weak, but is there any strong option out there?
Maybe. The proven marketing approach, which might only nod to the truth, is to “superset” a claim someone else has made. Sure, you say, AI might increase traffic, but you need to plan your networks for the Big Picture, of which AI is only a currently popular piece. I used to call this sort of thing a “marketing fable,” not unlike the old folk tale of the old woman who lived in a shoe. The lesson justifies the use of a colorful but obviously figurative example.
Or, if you like, there are some actual truths. IoT is one. Even leaving aside the devices that are part of smart homes and not enterprises, the volume of messages created here is already enormous; if I use enterprise data I get, grouped by verticals and then extrapolated on a weighted basis to the global market, then enterprise IoT generates more messages in a week than the whole of current IT processes in a year. If we harnessed this information to run a business more effectively, it would boost network traffic considerably, and analyzing and integrating it would also be a great mission for AI.
But you can’t just say “IoT!” these days and expect people to fall at your feet; been there, done that. Can it still be made exciting? I think so, but I admit that being fettered by reality makes things a lot more complicated, and we do have real IoT out there to limit our flights of fancy.
The point is that networking has lost its way to marketing and needs to find it again. Spec sheets are not marketing, they’re sales. Press releases aren’t marketing, they’re just ways of conveying a marketing message. Have network vendors forgotten the real marketing, the generation of interest and excitement? That would be understandable if all your buyers are doing is trying to get more for less. But giving up marketing surrenders control of strategy because only marketing can communicate strategy. Sales executes on it; marketing makes believers to sell to. If you want to push something new, you need believers to have prospects and sales.
That, in my personal view, is what will frame network competition beyond 2025.
Next up: My thoughts on ending the great network depression.