Google Compute Engine – Should Amazon be Worried?

I have been following the blogosphere and Twitterati ever since the Google IaaS announcement came out officially. Within 24 hours, there have been umpteen number of articles, posts and tweets about how Google is all set to take Amazon head on! Almost 90% of the articles mentioned that Google is ready to lock horns with Amazon. But really, should Amazon be worried? No! Not at all!

Amazon Web Services is ahead of any Cloud offering by leaps and bounds. The sheer breadth of services that AWS offers is mindboggling. The API that AWS designed for each of its service is considered to be an industry standard so much so that even the competition is left with no choice but to endorse it. Google’s Cloud Storage API is fully compatible with Amazon S3 API. Whether it is asynchronous messaging for loosely coupled components, a high performance SSD based NoSQL database or a full text search service, AWS has got it all covered for you! From an early stage startup running a simple website with a database to a large enterprise crunching their Big Data on the Cloud, AWS is the preferred platform. And, mind you, Amazon didn’t get there overnight. It took a good six years and abundant conviction that got them where they are today. AWS is built on the same philosophy of – customer centric innovation. Have you ever seen any executive of Amazon taking a dig at the competition? In fact, while Google was busy enjoying all the media attention like a teenage celebrity, the leadership team at AWS would be prioritizing the next key feature to be added to EC2. Jeff Barr, Amazon’s Chief Evangelist and the face of AWS would quietly post a new blog that would pleasantly surprise the customers and send yet another shocker to the competition. Amazon doesn’t care about competition as much as they care about customers and that is one of the reasons for their success.

Amazon’s dominance in the Cloud is acknowledged by one of their latest competitors – HP. Though HP now has a Public Cloud that aims to grab the market share from AWS, they were left with no choice but to support AWS. Apart from HP Cloud and Savvis, HP’s CloudSystem appliance can burst the workloads into Amazon’s Cloud.

While everyone is excited about the Compute Engine announcement, there is nothing new to write home about. It is yet another VM as a service on the Cloud. In the current form, it looks what Amazon EC2 was in 2008. Calling GCE as the Amazon killer reflects the over enthusiasm of a section of media and industry that are intimidated by Amazon’s dominance.

Talking about customer apathy, the developer community has been pleading Google to add support for PHP to the Google App Engine for the last 3 years. This has been the single most asked request from the GAE customers. On the Google App Engine Forums, PHP language support is the largest thread with thousands of developers posting this request. On Jan 6, 2011, one of the Google employees posted the following – “I'm making this issue read-only. I think the points here have been made. There's no reason to email thousands of people every time someone says '+1'. There are no current plans to support PHP on App Engine. No one on this team is against the idea, and given unlimited resources, we would do it. At this time, bringing another language runtime to App Engine is unfeasible given the other goals we are trying to meet.” Since then Google did nothing to address that. They were busy fighting the Facebook threat initially with Wave, Buzz and now Google+.

I am personally not very excited about Google’s Compute Engine announcement and there is a reason for that. I was one of the biggest advocates of Google App Engine when it was released back in 2008. GAE was the first PaaS that I got my hands on. Right after I deployed a simple Python app on GAE, I realized the potential that a PaaS holds for the customers. I was very happy to see GAE adding Java support in 2009 and I thought Google got it right. After 4 years of being in existence, did GAE make a dent? Leaving aside the magic numbers of adoption that Google shared at the Google I/O, I have not seen any of my customers and potential customers from India betting their farm on App Engine. The snail paced innovation and half-hearted execution of Google App Engine has killed the enthusiasm of the community. Heroku has been having Google’s lunch for a long time and it was all over when AWS launched the free tier! As a disappointed GAE developer, I am very skeptical about Google’s commitment to the platform and wonder how serious is Google about the developers. Like most of the Google technologies, it lived in the preview mode for almost 3 years! Google App Engine saw the release only in 2011. Having moved out of the preview, Google didn’t get the pricing right which upset the customers. Though the GCE pricing is announced on the same day, I am curious to see how long it takes for Google Compute Engine to come out of the preview.

In the last couple of months, many industry players announced their IaaS service. It was HP that went into public beta last month and Windows Azure that announced the VMs earlier this month. Oracle and Red Hat also announced their Cloud strategy this month. With existing IaaS players like Rackspace, GoGrid, Joyent, Verizon and AT&T and the new entrants like HP, Microsoft, Red Hat and Oracle offering exactly the same VMs on demand, where is the differentiating factor? Would startups, SMEs and enterprises flock to Google just because they let you run your applications on VMs? So, who has an advantage in this crowded market? The answer is clear – whoever gets their Hybrid Cloud strategy right will be the winner. Customers are no more excited about getting rid of their servers and moving the applications to the Cloud. They need a secure, reliable and integrated strategy to connect their existing investments to the Cloud. And, this is where I will bet on Microsoft. Windows Azure’s Hybrid Cloud strategy is very compelling for the enterprises. Through the Virtual Network, Service Bus, SQL Data Sync and Windows Azure Active Directory Services, Microsoft has the most compelling story for the enterprises. For a long time AWS has downplayed the Private Cloud phenomenon to an extent that Dr. Werner Vogels, CTO of Amazon called the Private Cloud as the false Cloud. But if AWS wants to penetrate into the enterprise segment they need to demonstrate their commitment to the Hybrid Cloud. Of Course, Amazon Virtual Private Cloud is very robust and mature service. But lack of presence in enterprise is a huge setback for AWS to drive the enterprise adoption. With the growing threat from OpenStack, Amazon was forced to work with Eucalyptus to unify the API layer. Microsoft has an undue advantage in this space. There are a few factors that work in favor of Microsoft in the enterprise. Firstly, MS has decade long relationship with enterprises and became a credible player in that space. Second, through Hyper-V and System Center Virtual Machine Manager (SCVMM), Microsoft is aggressively competing with VMware to quickly become an alternative for the Private Cloud of large enterprises. Thirdly, with the existing services that support Hybrid Cloud and the homogenous Cloud infrastructure running both the Public Cloud (Windows Azure) and the Private Cloud (SCVMM), Microsoft enables easy VM migration and may even support Cloud Bursting in the longer term. From compute to storage to databases, Microsoft is evolving the same unified stack for both on-premise and the Cloud. This will certainly give Microsoft an edge over the competition.

In my opinion, Google will have to deal with Microsoft first before aiming at Amazon. Both started as PaaS players and eventually embraced the infrastructure. Till then, let’s admit the fact that Amazon is the Cloud leader!

- Janakiram MSV, Chief Editor,


Updates from around the world