With the massive size of the public cloud market, it’s reasonable to assume that there’s plenty of pie for each the top three vendors –Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure (Azure) – to get their fill.
But the truth is that they’re all battling to capture even larger slices. While this type of war has happened in other technology segments, this one is unique because the market is growing at 30-40 percent year-over-year.
Here are a few examples of the current ugly wars these vendors are waging against each other.
AWS is luring away Azure customers. Channel checks suggest that AWS is incentivizing clients to move their Windows workloads to Linux. The next step is to move their SQL Server workloads to other databases (e.g, PostgreSQL). Of course, it won’t stop there; there will be an entire migration strategy in place. And there have even been a few instances in which AWS has funded clients’ early PoCs for this migration along with the implementation partner.
Azure is pushing for AWS migration. It isn’t uncommon for many mid-sized implementation partners to make their client pitch solely on the fact that they can migrate AWS virtual instances to Azure and achieve 20-30 percent, or more, cost savings. It also isn’t uncommon for Microsoft to bundle a lot of its offerings, e.g., Office 365, to create an attractive commercial bundling for its broader cloud portfolio against AWS, which lacks an enterprise applications play.
GCP is pushing Kubernetes cloud and Anthos. GCP’s key argument against AWS and Azure is that they are both “legacy clouds.” The entire Kubernetes cloud platform story is becoming very interesting and relevant for clients. More so, for newer workloads, such as AI, Machine Learning, and Containers, GCP is pushing hard to take the lead.
Each of these vendors will continue to find new avenues to create trouble for each other. Given that Azure and GCP are starting from a low base, AWS has more to lose.
So, how will the cloud war play out? Three things will happen going forward.
The vendors have realized that clients can relatively easily move their IaaS, and even PaaS, offerings to another cloud. Therefore, they’ll push to make their clients adopt native platform offerings that cannot be easily ported to different clouds (e.g., serverless). While some of the workloads will be interoperable across other clouds, parts will run only on one cloud vendor’s stack.
While the vendors will acknowledge that implementation partners will always have cloud alliances, they’ll push to have preferred partner status for specific workloads such as database lift and shift, IoT, and AI. For this, most cloud vendors will partner with strategy consulting firms and implementation partners to shape enterprises’ board room agenda.
In 2018, Google acquired cloud migration specialist Velostrata. This year, both AWS and Azure launched migration kits targeting each other’s clients. This battle will soon become even fiercer, and will encompass not only lift and shift VM migration, but also workloads such as database instances, DevOps pipelines, application run time, and even applications.
With the cloud giants at war, enterprises need to be cautious of where to place their bets. They need to realize that working with cloud vendors will become increasingly complex, because it’s not only about the offerings portfolio but also the engagement model.
Here are three things enterprises should focus on:
The cloud wars have just begun, and will become uglier going forward. The cloud vendors’ deep pockets, technological capabilities, myriad offerings, and sway over the market are making their rivalries different than anything your business has experienced in the past. This time, you need to be better prepared.
What do you think about the cloud wars? Please write to me at [email protected].
At this point in time, every CIO knows that the cloud is a more agile, faster and cheaper place to be. But they have a tremendous amount of technical debt in legacy ecosystems, which, so far, has proved to be largely resistant to moving into cloud. There are a lot of companies, like AWS, that will help you — for free — evaluate your applications portfolio and help you calibrate which applications can move to the cloud. So it’s not that we lack the knowledge or the technology to migrate to the cloud. The problem is that the people that manage legacy ecosystems resist their moving to the cloud.
Organizations are passive-aggressive toward change. By that I mean that everyone gives lip service to migrating to the cloud, but they see their job as telling the CIO the problems you’re going to face (security or performance, for instance) and why you can’t do whatever action you want to do. They will eagerly tell you that you’ll have problems, you can’t do it, or it will be very expensive.
The services market is aflutter with the coming impact of migrating work from legacy environments to cloud environments. Twenty-five to 30 percent of capacity in legacy environments today is utilized for testing, making it ideal as the first workload in services to migrate to cloud. We see signs that this migration is already underway. We believe the implications of migrating the testing workload are very significant.
Let’s look first at why testing will be the first big area to move to the cloud. Testing environments are low risk to be moved, and there are few production consequences. Testing is a really interesting first step and low-risk way to test cloud environments. Test environments are used intermittently, which makes them ideal workloads to move into a public cloud where you only pay for what you use. You can spin them up quickly and shut them down quickly. There is also a substantial cost benefit. And they’re a great test bed in that you can use them not only for the testing workload but also test your legacy applications running in these environments.
It will take some time to put the migration in place, but we think this is a natural grouping of things to move.
The implications, though, are fairly dramatic to the existing legacy workloads.
If a testing environment comprises 25 to 30 percent of the capacity moved, it means you’re left with excess capacity in your legacy environment. For organizations that are not in outsourcing relationships for testing, they won’t need to buy new equipment; the capacity can be used for production as production grows. But those that are in outsourced relationships, whether with asset-intensive or remote infrastructure management outsourcing (RIMO) providers, there will be a corresponding reduction in the invoice to the outsourcer.
At Everest Group, we’ve looked at the contracts for more than 300 outsourcing relationships to see what the implications would be as a result of the migration. We found that almost uniformly, if companies manage these migrations correctly, they can reduce their invoice by 25 to 30 percent without breaking banding or costing prices. So testing workloads certainly can be migrated.
This will result in a downward pressure in revenue for the incumbent infrastructure providers, so we think this has significant implications for the outsourcing industry. Whether it is the RIMO players or the asset-intensive players such as CSC-HP and IBM, migrating the testing workloads will impact 25 to 30 percent of the existing capacity in these providers’ invoices.