A modest cloud proposal
This week we had the most recent Gartner IaaS magic quadrant confirming that cloud is a 2–3 horse race between Amazon Web Services (AWS), Microsoft Azure and (maybe) Google Cloud Platform (GCP). The WSJ also reported that Rackspace is about to be acquired by a private equity firm.
The obvious economies of scale and the movement of the magic quadrant over the past few years indicates to me that there is only one strategy in public cloud: go big or go home.
So if I were a private equity firm, that paid several billion dollars to buy Rackspace, what would I do?
I posit that for every AWS service there exists a reasonably competitive alternative but that nobody has tried to put them under one roof to get the benefits of scale and integration that AWS provides (along with GCP and Azure).
Lets go through some examples, starting with the foundational AWS services:
Compute, EC2 competitor: Digital Ocean. Estimated to be the second largest hosting company in the world. Projecting from Netcraft’s 2015 server count and Digital Ocean’s growth trajectory I estimate they now have around 350k web servers (around 3x what Rackspace has). If the average selling price is $10 per month works out to a $42 mm run rate. This is a pittance compared to the revenues of AWS or Rackspace but growth, developer mindshare and scale make it a worthwhile investment.
Storage, S3 competitor: Backblaze. Now that Backblaze have launched B2 and are storing over 250 PB of data they are a serious competitor to S3. Pricing 250 PB at rates similar to S3 of $0.01 per GB per month gets you a $30 mm run rate. I estimate Backblaze to be growing at 50% annually based on the time between them reporting 100 PB under storage (March 2014) and 250 PB (July 2016).
NoSQL database, DynamoDB competitor: mLab Only $7 MM in revenue. But hey, again, growth and mindshare. Everybody uses Mongo right?
SQL database, RDS competitor: Heroku Postgres?? (OK I’ll admit this is a stretch).
Notifications, SNS competitor: Twilio. $166 MM revenue reported in their S1 is not too shabby (unfortunately for my grand plan Twilio is now public).
Other longshots: Akamai (1.6B revenue), Heroku (owned by Salesforce), Github & Atlassian.
There are also lots of smaller companies like TravisCI and PagerDuty that might also help to round out the full suite of applications.
Combining the above companies with Rackspace would give you synergies (e.g. putting backblaze pods in Rackspace DCs), several billion in revenue and competitive offerings in the most important cloud service segments. You will have about $2 billion in revenue, equivalent to AWS in 2012. You will probably have more revenue than GCP and be in the same ball park as Azure.
How much would all of this cost? Some of these companies are public and have several billion dollar price tags (e.g. Twilio). But, they are all smaller than Rackspace. Plus, you will need a war chest for the inevitable price war with 3 of the most cash rich companies on the planet. Anybody got $20 billion dollars they want to lend me?
Why do this?
There are strategic advantages to being an independent cloud services provider. Walmart (for instance) won’t want to run their cloud on AWS because they compete in retail. Google likewise is a legitimate threat for many businesses, especially internet businesses (think Yelp). Microsoft might seem more benevolent to the average enterprise until they bought LinkedIn.
Does anybody else thing somebody is going to try this strategy?