AWS re:Invent takeaways…
- The borders around the Hybrid Tech market are likely to erode further in the coming years as AWS took many steps to make workload migration easier, run the platform cheaper for clients, and continue to add new services and make the platform even more compelling
- AWS innovation curve is so rapid and ubiquitous across the software spectrum that it continues to gobble up new markets away from everyone else
- Some AWS relative weak spots were fixed with improved functionality, diminishing some opportunity for IaaS peers
- The innovation curve includes re-writing AWS code to make features/functions faster which is not unlike a price cut for existing customers to stimulate elasticity
The “Why” & “How”:
- Taking the leap to Serverless
- Making it even easier for migration and ingestion
- New launches include Cloud9 for collaborative programming, SageMaker for ML, Rekognition for machine vision learning, managed Kubernetes as a service, advanced load balancing functions (ALB, NLB)
Good anecdotes that exemplify the problem facing the IT world:
- Confusion: partner or competitor. Companies spent top dollar booking Expo booths as part of the AWS ecosystem, accustomed to a two-way sharing of customer funnel with the AWS platform. These companies hoped to engage passersby on the show floor as potential customers. But when AWS launched competing product at either of the major Keynotes, those companies spent the day answering the question of how they would survive an AWS threat.
- Do they want it all? By the end of the conference we had similar conversations with a large enterprise (currently in the vicinity of 25% in public cloud with plans to go to 45%), a global colocation company, and a global OEM integrator. In one form or another, they all wistfully joked / shrugged / wondered just how much of IT does AWS want to own…with the not so subtle implication that only AWS could limit its eventual dominance.
Some implications for specific stocks:
- SPLK: inflection in the need for machine log data, this was evident in AWS sessions on analytics and database products, and also through our candid discussions with SPLK competitors
- RHT: increasingly easy migration tools from AWS and maturing workload mobility tools (containers) limit the opportunity for OpenShift, and core RHEL on AWS is unlikely a long term match. A 5-year view on RHT looks fairly bearish, from the current set of product and competitive landscape. We kind of walked away thinking – what will RHT do for a living in 5 years?
- VMW: we thought it would take longer for the AWS trapdoor to be evident, but it was on full display at the conference where VMW on AWS featured as an expensive and maybe unnecessary migration tool for workloads to eventually shut down VM rich on premise production.
- ORCL: many of AWS’s weapons seem to be trained on ORCL, as the company was held up as poster child for bad behavior among software companies, and new AWS service launches seem increasingly geared at dismantling ORCL’s hold on customer database demand.
- AYX: sophisticated tool, important category, limited competition today, and somewhat fervent user base were positives. Negatives include limitations (today) around non-relational data, and long term negative could surface if ‘data preparation’ functionality is in the glide path for future AWS SageMaker launches.
- AKAM: AKAM’s best product, Kona Site Defender, was reviewed in depth for its weaknesses relative to AWS Shield. Put simply, companies who shift to AWS Shield don’t experience DDOS attacks any more, and don’t have to worry about slow response times for help from the Kona team (up to 8 hours), or 24-36 hours to get the logs. It makes sense that AKAM is now focused on alternative areas of the enterprise security landscape.
We have written about the theme of 'hybrid hype'. Specifically, in the last 18 months we have seen many legacy enterprise equities rally around the theme that ‘not all workloads belong in the cloud’. What we saw at AWS re:Invent this last week further confirmed to us that “not all” is a shrinking # with an apparently limited future.
