OK, time to look at some radical news. HPE is said to be acquiring Juniper. If you’ve followed my writings, in my blogs and in articles I’ve done for tech publications, you know that I’ve been talking about the shift in how network budgets are influenced. There was a time, perhaps only a decade ago, when people bought network gear and engaged with service providers based on network-technical factors. However, as far back as 25 years ago, the leading influences on network planning were most often computer vendors like IBM. Today, computer/software vendors are the major influence in networking, period. So, building on a popular characterization, the dog is finally taking full control of the tail.
Rumors of this came out late yesterday, and it’s been picked up by the WSJ and other financial sources, in addition to the Reuters piece I’ve cited here. I had emails from three Wall Street types and two other sources this morning when I checked my mail the first time. The story I hear is that the deal is likely to be announced later this week, and that’s also cited online. This isn’t a sure thing, but let’s assume it really will happen and examine how the deal relates to things like that shift in network influence I’ve talked about.
Way back in the 1980, buyers overwhelmingly cited data center technology sources as the prime strategic influences in networking. IBM, in fact, had as much influence (I used a scale of 0-100 to reflect no impact on decisions up to decisive impact) as every other source combined. The reason is obvious; the business mission of networks was tied explicitly to the applications that supported business operations. What the applications needed was what the network had to provide. However, over the next twenty years or so, the business challenge was empowering workers with information that their applications and data resources already had. Thus, network technology relating to worker connectivity was the hot button. However, even as this trend was developing, there was already an underlying shift underway, an increased focus of the “network budget” not on supporting new things but on sustaining what was already there.
Where we are today is that networking is part commodity and part slave to new IT directions that could open new benefits and justify new spending. Network equipment vendors have no real ability to open new business cases because they require new software spending, and the vendors don’t write software. One comment I made recently was that it was interesting that Cisco, who has the most IT assets of any network vendor, has somehow failed to capitalize on the way those assets could drive further network project, and spending, growth. What I think is going on here is that HPE wants to do what Cisco has not.
Why now? Part of the reason is that if I can see the trend, it’s unlikely that vendors on both the network side and the IT side can’t see it. Cisco’s IT assets are a kind of hangover threat, because Cisco is a sales monster and if they really decided to get IT-strategic they could make life miserable for competitors. IBM has gained significant traction with the acquisition of Red Hat and their populist story of hybrid cloud, and while they sold off their own network business years ago, they could easily tap into the open-white-box model and enter the space again as an integrator. Dell has network gear, and so does HPE, but neither is a big name in the network space. Juniper is a path to becoming one.
Another reason relates to another trend I’ve been blogging about. The notion of “best of breed” is falling by the wayside because enterprises are finding it increasingly difficult to integrate a series of targeted technologies to create a full network/IT ecosystem. Finger-pointing and fault-isolation problems are just too difficult to resolve, and so there’s been a clear trend over the last five years toward the one-stop shop. And if you want to be a one-stop shop, one thing you darn sure have to offer is the key element that probably drives the unified deals. That’s not network equipment, so if you are a network vendor you have to wonder how you hold a place in the one-stop future. Answer, you have to be more than a network vendor, which for most means you have to be acquired by an IT player.
We can also link this deal with the recent Broadcom acquisition of VMware. It can’t be a coincidence that we have two big tech M&A deals in short order. What’s happening with Broadcom is a recognition that you can’t drive chip sales in a vacuum, that you need to have something that will fulfill a broader mission so that you can generate enough value to make a place for yourself in the market. Bigger is better, both for Broadcom and for HPE.
Why Juniper, as an HPE target? The Street and most of the tech press are going to make this about Juniper’s Mist AI and AI portfolio overall, and I think that’s a part of the reason, but to suggest, as my cited article suggests, that “The deal would help bolster the nearly 100-year-old technology company’s artificial intelligence (AI) offerings, according to a WSJ report earlier in the day.” invites the notion that HPE wants to buy Juniper for AI, and other financial sources make that point even more directly. I think that Juniper’s AIOps story would be highly valuable to HPE as a means of unifying at the operational level the IT and network domains that are already unified now at a mission level. But they could achieve this without buying a full-spectrum network company. No, this is about making the “shop” big enough to be a “one-stop shop.”
The combination of HPE and Juniper could be an industry game-changer, because there is no other player who offers a complete server/software IT solution and a complete network solution. The opportunities to exploit that dualistic position at a time when it’s what buyers not only want but need, in order to advance the role of IT overall, are dazzling. If HPE is number two in computing and Juniper number two in networking, the combination is Number One.
If they can make it work. Any kind of deal like this is going to depend on one fundamental force, and that force is marketing. I’ve always said that Juniper had the best technology overall in the network space, but that their marketing and positioning didn’t live up to its potential. You could say the same thing about HPE. There is no way that you can create a marketing monster by adding two marketing children together. What you get is a group of children. Somebody has to pick up the role of adult here, and that somebody has to be HPE for two reasons. First, HPE is the acquiring company so they will prevail in a management sense. Second, HPE owns the IT missions that will have to build the new business cases and drive new combined IT/network projects. But can HPE do it? Candidly, I don’t know.
This deal, if it’s real, will rival the Broadcom/VMware deal in terms of influence. It will further subduct networking under the forward, if tectonic-paced, evolution of application software and software platforms. The network will be, for enterprises and consumers, even less visible. But that was happening anyway. This is an example of how companies will try to deal with the changes that the market are already forcing on us all. Let’s see if it’s a good example.
