Technology Predictions for 2019
I used to write an annual blog post on my networking predictions for the year ahead, but it seems that I dropped the ball (as it were) at the end of 2017. My last effort — predicting the rise of SD-WAN, expansion of network virtualization outside the data center, and the importance of networking for clouds and containers — is holding up pretty well I would say. Now that I’m almost two years into my role as APJ CTO for VMware — with a remit that covers all of IT, not just networking — I’ve dusted off the crystal ball for 2019.
As I mentioned in some of my talks last year, I was pretty convinced that Artificial Intelligence was just about to take off when I started my Ph.D. in 1985 — so I was only about 30 years too early with that one (fortuitously I dodged the AI Winter of the late 1980s by doing my Ph.D. on formal methods). But at this point we’re clearly seeing AI (and the related field of machine learning) truly become mainstream. In 2019, we are sure to see more and more applications of AI and ML across all sorts of industries, including applications within VMware’s products. At the same time there is reasonable concern about the ethical and social implications of allowing machines to make judgements — an issue that pre-dates AI but which is going to get more attention as AI spreads.
It is clear that successful application of machine learning algorithms depends on high quality data. So while ML algorithms might be viewed increasingly as a commodity, many enterprises have unique data sets on which algorithms can be trained, and this provides a path to differentiation.
It doesn’t take much foresight to see cloud adoption continuing to gather steam, but two cloud-related trends stand out to me:
- Enterprises have generally found adoption of public clouds proceeding more slowly than they hoped, due to a broad set of issues ranging from security to application architecture.
- Most enterprises find themselves in multiple clouds, both public and private, and that is likely to persist for the foreseeable future.
This points to the increasing importance of technologies that make it easy to operate across multiple clouds. These technologies also need to address the needs of multiple constituencies: developers, line-of-business owners, central IT, site reliability engineering (SRE) teams, and so on. We will continue to see plenty of investment and innovation in technologies and tools to help enterprises manage the multi-cloud world.
Security continues to be an issue with board-level visibility and cyberattacks are only going to get more sophisticated. We should see an increasing focus on good cyber-hygiene — and I would like to think that micro-segmentation will join the list of things that sits in the cyber-hygiene category. We should view micro-segmentation as one of the basic things to get right, just like patching and multi-factor authentication. At the same time, a fundamental re-architecture of infrastructure to improve security — as advocated by my colleague Tom Corn — is within reach.
IoT and Edge Computing
IoT is another technology whose benefits we’ve been looking forward to for a long time (the first Internet-connected soft drink machine seems to date from the early 1980s). With lots of IoT pilots underway, we have to expect increasing adoption of IoT, although I would say we are well past the peak of the hype cycle now. For me, the most important IoT development of 2018 was the widespread acceptance that Edge computing is a real thing (notably with the Outposts announcement from Amazon). So 2019 should see Edge computing really taking off (I can claim a small contribution to predicting the rise of the Edge in 2005, but the main credit for farsightedness belongs to my co-authors).
Finally, I have to mention Blockchain, which, like IoT, is arguably past the peak of the hype cycle. I remain optimistic that the technology has potential — especially outside the world of cryptocurrencies. Professor Emin Gün Sirer of Cornell gave my favorite technical talk of 2018 (at future:net) and I was persuaded by his argument that even if every cryptocurrency fails, the techniques that have been developed to build distributed ledgers will have enduring value. 2019, I predict, is the year that enterprises will start to find that value by experimenting with real blockchain use cases at scale. Hopefully the path to successful deployments is faster than that for AI.
You can watch me take a look back at the tech trends of 2018 and discuss some of the predictions covered above here.
Photo by Bruce Davie: the penultimate sunset of 2018, Falls Creek, Australia.