We’re almost done with 2022, a year that a lot of people aren’t going to be sad to leave behind. The big question for us all is whether 2023 will be any better, and there are signs that point both upward and downward. The tech world is always a slave to the broad economy, but the two don’t move in lockstep so we have to look a bit deeper at the trend lines. To do that, I’ll draw on what I’ve learned from both buyers and vendors, and also on Street research.
Keep in mind that both these sources have potential issues. I don’t necessarily see a valid statistical sampling of viewpoints, because my contacts come from people who are likely following my views. Whether I influence them, or simply “pre-select” from the broad market because people who agree with my views are more likely to contribute to my information, there’s a selection bias at play. With regard to the Street, I’ve noted many times that there are things they’re good at (like whether budgets are going up or down) and things they’re terrible at (like almost everything relating to technology insights). With that in mind…
If we start with the enterprise, we see a convergence of views from both sources. My contacts tell me that budgets for 2023 are not under pressure, but it is possible that some projects will be pushed a bit toward the second half because of broader economic uncertainty. The Street view with regards to budgets is similar, but I don’t see a significant consensus on the project timing issue.
What is also consistent is the general focus of projects. There’s a lot of interest in vendor consolidation, which I see primarily at the hardware and platform software level but which the Street sees as a broader trend. The issue here is that enterprises think that integration of any sort has become a major problem, and they’re looking to reduce it by reducing the population of things that need to be integrated.
One insight I got from my contacts that the Street didn’t mention is that there’s a movement toward “full-stack” or suite vendors and away from the “best-of-breed” thinking. I think this will favor vendors like IBM/Red Hat, VMware, Dell, and HPE. All these players offer both hardware and platform software, and that’s looking increasingly attractive to enterprises. Part of the reason is that enterprises are reporting rapidly mounting difficulties in acquiring and retaining highly qualified technical personnel.
The highest priority for enterprises is the applications, though, not what they run on. There is still significant interest in using the cloud to improve engagement with customers and prospects, and also (increasingly) to support employees and suppliers. There’s also growing interest in using SaaS for horizontal applications, and in fact SaaS use is the largest reason for “multi-cloud”.
Things on the network side are significantly more muted. There is no budget increase on tap for network infrastructure among enterprises, only for security products. This is another area where my data matches that of Wall Street, and it’s hardly surprising, given that the role of the network isn’t changing and so major transformational investment would be hard to get past the CFO.
“Transformational investment” is a worthy topic in itself. My data has consistently shown that in both networking and IT, there’s less interest among enterprises in trying out new stuff than we’ve seen for at least a decade (in fact, since the downturn in 2008). New technologies, new vendors, new anything, is kind of out of favor, even with CIOs who are usually more interested in that sort of things than the rest of the C-suite. The Street has generally picked up on this point, too.
What the Street doesn’t validate specifically is the attitude enterprises have toward even identifying new technologies, and perhaps surprisingly, toward discovering new paradigms for the way that IT is applied to their businesses. “Stay the course” is the mindset, and part of the reason I’ve heard for this is that buyers have lost faith in many aspects of tech. The number who say that they trust the representations made by their current or prospective vendors has dropped by a third in the last five years. The number who trust media and advertising has been cut in half over the same period.
In the network operator space, things are way more complicated in some ways, and much simpler in others. Complicated because there are many different types of operators, and many different sized. Complicated because an operator organization is like a government; huge, monolithic in some ways and yet home to radically divided opinions. I described the latter phenomena as the young Turks versus the entrenched conservative management establishment, but in fact you could divide operator organizations up in three or four different ways, which shows just how hard it is to establish a perspective on “plans”.
Let’s cut through that, though, but admitting that senior management falls into the “entrenched and conservative” part of any division, and that this group is still driving the bus. If we do that, we can see some surprising congruencies with enterprise management.
The Stay the Course mindset is one. It may be that management thinks that all the current tech market confusion is something that will pass away without any need for them to act in any particular way. It may be that they don’t think that they can sit on their hands, but don’t know what to do when they stand up. Understandable, when you have asset useful-life expectations that are on the average from 50% to 100% longer than you’d see for the same assets in an enterprise deployment.
This attitude is prevalent in both vendor and technology selection, but in both areas there’s a bit more of a split view among operators. There are some who are eager to change both vendors and technologies, in order to better manage costs, but for most this attitude is confined to areas of the network (geographic or service segments) that are largely greenfield or are undergoing major modernization. 5G is a good example.
Cost management, in fact, is the only consistent driver I see among operators I’ve chatted with. There is really little interest in “new services” beyond making some largely inconsequential tweaks to existing services and calling the result “new”. That’s focused operators more and more on the cost side, but at the same time ensured a collision between their goal (reduce costs) and their technology intransigence (don’t change vendors or technologies radically). Part of that, of course, can be attributed to the long capital cycles operators have to deal with, but part is just the traditionalist bias of senior management.
There are some things we can expect to see changing in 2023, though, but in a more evolutionary way. For consumer services, we’re recognizing that faster/better Internet sells, even if the claims are vague. As a result, there’s interest in pushing fiber broadband further, even if it means that there are little broadband enclaves where demand density is high, surrounded by areas where broadband is relatively poor. There’s also interest, and rapidly growing interest, in using 5G in general and millimeter wave in particular, as an alternative to fiber/wireline.
Consumer broadband, the Internet, and the cloud, are now linked in a powerful symbiosis. Over the last two years, we’ve seen the value of engaging users online to promote, sell, and support products. That has led to a demand for richer online experiences, which has led to both a need for cloud-hosted front-end elements to applications and better broadband to deliver richer visualizations. As a result of this, business networking is being transformed, in a multi-dimensional sort of way.
One obvious dimension is that better Internet and consumer broadband will make better low-cost connectivity available in more and more areas where there’s a population concentration that raises demand density. Since those are the same areas where business sites are likely to be located, it’s pulling businesses toward Internet exploitation rather than specialized business Ethernet/fiber access and MPLS VPNs. Major sites aren’t likely to end up on consumer-driven broadband infrastructure, but any smaller sites will be better supported there, and so the migration of small sites toward consumer broadband is almost assured.
The other dimension is the cloud/Internet role in supporting application front-ends. If customers, partners, and (increasingly) workers are all supported via the Internet and the cloud, then “cloud networking” or the connection of data center sites with the cloud for hybridization, becomes the primary incremental enterprise network requirement. Even alternative VPN technologies like SD-WAN are more and more likely to be deployed as virtual-network SASE instances in the cloud than as branch appliances or services. If workers access their business applications through the cloud, then all you need in branch locations is an Internet connection and security tools.
All this adds up to the need to create a unified virtual-network model for enterprises. I think that the announcement made today by Red Hat and Juniper is an example of that; Juniper has both SD-WAN and general virtual-network technology (from their acquisition of 128 Technology and via Contrail, respectively), and the paper posted today is an example of how we might see networking transformed by the combination of the cloud, the Internet, and even consumer broadband.
Security is the final point here. We have been practicing incrementalism with respect to security, like a crowd of children with fingers at the ready to stick in holes in our dikes. It’s time we fix the leaks holistically, but that’s been difficult to do given that the evolution of the cloud front-end hybrid model has created new attack surfaces, that network connection options are evolving, and that vendors tend to introduce band-aid fixes rather than asking buyers to start their security considerations over again.
Years ago, I pointed out that security really has to start with session awareness, meaning that network services that can identify a specific user-to-application relationship and match it against what is permitted and what isn’t, should form the foundation of a security model. We are starting to see that, and starting to understand that some form of virtual networking, which elevates connection management above transport connectivity, is essential. For business services, then, virtual networking is the most important thing to consider in 2023.
We’re not going to see a network revolution, or a tech revolution overall, in 2023. Economic concerns and central bank tightening will ensure that growth questions will haunt some economic sectors through the year, and that will make both businesses and consumers a bit more cautious. What we will see, I think, is a number of evolutionary changes in our tech ecosystem converging to create some trends that are a bit more visible, and thus perhaps more addressable.
I’ve tried this year to present an objective view of tech, offending some along the way and perhaps exciting others a bit. I’m pledged to do the same thing in 2023. This is my last blog of 2022; I’ll resume after the first of the year. Meanwhile, have a Happy Holiday season!