WP004 - Gartner’s Hype Cycle for Platform as a Service, 2017

This week, we look at Gartner’s recently published PaaS hype-cycle. There’s a lot of categories in it and PaaS features, making it a huge document with many issues around “is that really a thing”? The answer is that, for the most part, yes. Brandon and I look at the structure of the hype cycle, how different audiences should use it, and even the content itself.

Our notes this week are a little light, apologies. As Brandon said, “this was quite the document.” Brandon.

So, Gartner…

  • Yefim V. Natis, Paul Vincent - this is the PaaS team, they work on the MQ and other things.
  • It’s best to treat this as an encyclopedia, don’t read it straight through, just what you want. What’s important about these documents is the format of it, it’s regularty, and therefore it’s long-term use for vendors and EA-types. You can argue about the ivory tower nature of it, but Gartner analysts don’t play: they follow a methodology that’s more trustworthy that YOLO or the impossible task of actually using all these things and following up with production users. The alternative is: do nothing.

The utility of hype cycles

  • Hype-cycles, how should they be used? Pacing, and a survey of the space.
  • Doing “predictions” is weird and strange. You have to come up with a method and justification. I always tried to avoid them, or make them absurd.
  • I’d use these in strategy and M&A work to estimate “marker-windows.” Again, it was better than nothing. That said, a spread of “5% to 20% of target audience” is pretty useless.
  • You do see mistakes in these, for sure, esp. as the analysts try to figure out how to categorize things. As this one notes, microservices is no longer tracked here, but another hype-cycle. On the other hand, it’s hard to imagine how microservices would work well without a PaaS in place, even it’s the “no one dare call it PaaS” of kubernetes plus something like Istio. “Container management” will probably moved out of here in the next few years.
  • Keep in mind that penetration is not 1:1 with the geography of the curve, like, at all. To get to the plateau something only needs 20% adoption: “Approximately 20% of the technology's target audience has adopted or is adopting the technology as it enters [Plateau of Productivity].”

The content

  • “The majority of PaaS-related technologies will reach or approach maturity in the next five years.”
  • Do we really need “Serverless PaaS” and “fPaaS”?
  • “etalon”! - “a device consisting of two reflecting glass plates, employed for measuring small differences in the wavelength of light using the interference it produces.”
  • Meanwhile: “The dubious viability of some pure-play providers, especially in the light of an inevitable consolidation of the market that will take place over the next three to five years.”
  • Lock-in: "users are advised to make an effort to minimize lock-in with any vendor or service offering in order to afford themselves an option of change in the future as the vendor strategies and user requirements inevitably continue to change” - this market is early, fear of lock-in.
  • The “PaaS” category inside Gartner.
  • “Private PaaS may be established on-premises or in a remote data center” - you have no idea how long it’s taken to get this…
  • “As more IT leaders have gained experience of the limitations of cloud-inspired technology in terms of delivering cloud-native outcomes, disappointment has grown, which is pushing private PaaS deep into the Trough of Disillusionment.”
  • Harshin’ your private cloud buzz like a boss: “Where possible, choose public cloud services. They should recognize that establishing a private PaaS is not a trivial exercise — it will require a lot more effort, skill and experience than subscribing to a public PaaS service.”

Past hype-cycles

Just for funsies!
2017: 31 dots
2016: 32 dots
2015: 31
A related domain, from back in 2011, “cloud computing”: