Last week Amazon published financials for its Amazon Web Services (AWS) business for the first time. These showed that the old man of public cloud computing is a $5bn business with 49% year-over-year growth. The data also showed that Amazon experiences healthy margins on AWS which puts them in a good position to handle the ongoing price war in the public cloud market. AWS also remains the leader in terms of cloud platform market share, ahead of nearest rival Microsoft by a good stretch.
The market was excited by this data as it not only shone a light on Amazon’s cloud business but allowed many to draw conclusions about the competition, primarily Microsoft, IBM and Google. What has been interesting in the days since the announcement is some of the coverage. There seems to be a belief in some quarters that the definition of cloud is Infrastructure as a Service (IaaS).
Various articles have compared AWS, Microsoft, IBM and Google’s revenue numbers with respect to cloud. For example IBM also released numbers last week reporting a $7.7bn cloud business with 75% growth in the quarter. However, a frequent conclusion to this is that the others are distorting their figures by including Platform as a Service (PaaS), Software as a Service (SaaS) and other such services. Some even went as far as to criticize Microsoft and IBM for including private or hybrid cloud related income, where they are deriving income from customers’ on-premise data centres. This is all bizarre because cloud is clearly more than just IaaS. To say that Microsoft should not include Office 365 as cloud income seems nonsensical.
Perhaps what these articles were reflecting was a frustration about not being able to compare like for like. for example, how does IBM IaaS revenue fare against AWS? But this again seems to misunderstand what cloud is. In a way it also misunderstands AWS which is not 100% IaaS either.
Apples and Oranges…It’s all fruit.
There was a time when cloud was clearer than it is today. Though let’s not forget that Amazon started life with a Queue Service, hardly the IaaS we know today. But in the last few years it has come to be seen by many as virtualisation of the data centre. Instead of applications running on physical boxes they run within Virtual Machines that allow far more efficient use of infrastructure (compute, memory, storage, power and so on). Within this world of cloud though we also have SaaS businesses like Salesforce.com, the quintessential cloud service. If IaaS means cloud then is Salesforce not a cloud business? Today many of us use other SaaS offerings such as Dropbox or even Netflix.
Then there is PaaS which while its success has been muted is for some more of a pure cloud platform for application development than IaaS. Finally a number of other cloud related services has emerged especially amongst large IT service companies such as Accenture or IBM. These focus on enterprise customers who are often challenged by moving from traditional data centre environments to either IaaS or PaaS.
But the story is still not that simple. With many organisations having concerns about issues like security or large legacy investments in software and hardware they do not all want to embrace public cloud. Instead they want to cloud enable their existing infrastructure and/or connect with public cloud. At the same time there are modern application architectures emerging that combine IaaS, PaaS and SaaS. Therefore to say that IBM and Microsoft should separate IaaS revenue from PaaS or public cloud from private and hybrid simply would not reflect the reality of customers modern requirements.
Why should it be that IBM helping a large enterprise to migrate some existing workloads into a cloud environment is less worthy of being classed as cloud revenue than Netflix migrating its entire stack to AWS? Just because one vendors cloud business is different in make-up to another should not make it less worthy of its revenues being counted as cloud.
What matters will still matter
The fact is that Amazon currently offers a certain collection of cloud based services which are largely IaaS and largely for development. Note that when I referred to AWS market share earlier I stated “cloud platform” and not cloud. With respect to that, it is doing good business both financially and in terms of market share. However, customers have many and diverse needs and other vendors are providing solutions for them: whether that’s Dropbox providing file sync and share or Google’s Nearline for disaster recovery.
If in the end, these vendors end up being almost 100% cloud based, then we can maybe look to slice and dice the revenue. But essentially what will still matter to those who care about numbers is which business is seeing the best revenues, margins and growth. Maybe that will mean that Amazon will sell more IaaS than Microsoft or IBM, but then Microsoft and IBM could well sell more SaaS. Either way, it’s all cloud!