The most recent Gartner report for IaaS, July 2019 – Infrastructure as a Service pictures a very clear situation in terms of leadership:
- AWS commands in the Virtual Machine market
- Azure comes in as the healthy second and will keep aiming for the top stop and might eventually get it
- Google with its GCP it’s a distant but a consistent third
- There’s a total of six (6) vendors on the analysis 5 of which are US-based. Alibaba is the only non-US organization
Overall the IaaS market is very healthy and projected to grow to $81.5 USD Billion in 2022 up from $41.4 Billion in 2009, almost a 100% increase.
The highlights of the report:
On the top three. The leader’s Quadrant
- Very well-entrenched and embedded in the enterprise.
- Organization of all sizes trust AWS as a platform because of its maturity and enterprise-grade.
- It normally is the first choice when talking about data centre migrations.
The not so good
- AWS aims to be the first on the market for new services and capabilities and because of it some of those services are very basic and/or lack deep platform/standards integration. I can remember Lamba Function and Elastic Bean Stack being in this category
- Amazon’s overall ambitions become a blocker for companies on the same vertical to use AWS. A clear example of these is Wal-mart in the US and Coles here in Australia.
- Almost every medium to large organization relies on Microsoft in its infrastructure services: Identity, Authentication, Monitoring, Software deployment, Server Operating System(s), Desktop Operating System. Organisations, in general, understand their ecosystem and it’s natural to consider Azure as the target Cloud
- The end-to-end integration starting with the development of .NET applications with Visual studio all the way to the deployment and integration with native services: Database, Storage, Integration, Alerting and Monitoring, etc, is unsurpassed
- A good integration vision to bring the best of class to the platform independently of the vendor or technology, Linux, for example, is the most run OS in Azure
The not so good
- Growing pains with the platform and its related reliability. In specific some Azure AD and because of the high dependency and criticality of the component, multiple services suffered downtime
- The over-promises made by Microsoft that moving to the cloud is ‘easy’ and ‘seamless’ and it could begin with a small investment as a headstart and at the end taking longer than expected and being very difficult. Microsoft’s internal sales organization can certainly oversell, often because they themselves don’t fully understand the whole offering nor the implications for the customers
- The poor quality of Microsoft’s own technical resources, support personnel and solution architects are mentioned in the report. I’ve had some very good experiences especially with technical support when solving critical problems and some very bad experiences with the field architects hearing them say completely inaccurate things
- A history of innovating technology which positively translates to its Cloud. I remember reading this article at the New Yorker, quite interesting and sometimes inspiring: https://www.newyorker.com/magazine/2018/12/10/the-friendship-that-made-google-huge
- The analytics and machine learning experience of Google as a whole makes it very appealing for enterprise customers
- Their Customer Reliability Engineering program which I’m not aware of but seems to be making a good differentiation for operations
The not so good
- Immature with Enterprise customers, this is true. I normally deal with large enterprises and Google is normally present but on very small and specific workloads nothing on the scales of Microsoft or Amazon
- A very small ecosystem of partners. This is key, they will not grow much of they don’t rely on a partner to sell, promote and support their platform
- Similar to Microsoft, solution architects on the field don’t seem to properly understand customer requirements
These three vendors are the most relevant, sure, Oracle and IBM have their own niche but they came very very late to the game. Oracle made IaaS an offering in 2016 and IBM acquired Softlayer around the same time.
Each of the top three vendors has their very specific and worth using strengths. I’d stick with AWS for EC2, bare-metal, RDS and many others that fit the requirement, in the case of Azure the strength is on SaaS and PaaS and they will keep betting on it. As for Google, I don’t think I’d recommend customers to move workloads to their IaaS layer as the first option mainly because the other two can offer a much better alternative. Google is strong for applications, data, analysis, global scale, that’s their muscle.