Larry Ellison takes on the Cloud – Entertaining but far from Accurate

October 5, 2009 Appirio

Balakrishna Narasimhan

The blogosphere is abuzz with the news of Larry Ellison’s hilarious evisceration of cloud computing at a recent event. As usual, Larry was funny, quotable and spinning the message to suit Oracle’s ends. Larry offered up a number of pronouncements about the cloud, which are worth tackling one-by-one.

“Cloud computing is not only the future, it’s the present and the entire past of computing”
Our assessment = 66% accurate
Larry is right that cloud computing is the present and the future of computing. We’re seeing companies like Flextronics, Dell, Cisco, Japan Post, JohnsonDiversey, Starbucks and many more adopt and enjoy tremendous success with the cloud today. Cloud computing has already come to the enterprise in the form of SaaS applications like salesforce.com and Google apps. In the future, as platforms-as-a-service get broadly adopted, more custom applications will move to the cloud and eventually we’ll see a future where companies cloudsource their entire technology foundations. Of course, this will not happen overnight but companies who experience the productivity and business benefits of the cloud want to accelerate rather than slow down the transition. The only ones incented to slow down the transition are vendors like Oracle whose fat 40%+ margins rely on the traditional software delivery and licensing model.

Where Larry is wrong is in saying that cloud computing is the past of computing. This is the tired argument that mainframes are akin to the present-day cloud. Mainframes did not provide a flexible, massively scalable and open platform to run your business. Even firms like Gartner, who rarely go out on a limb, say that cloud computing is the most significant shift in computing since the move to three-tiered architectures.

“Cloud, it’s databases, and micro-processors and memory and microprocessors and the internet”
Our assessment = 20% accurate
Larry is right about the cloud relying on traditional computing building blocks and internet standards. Unfortunately, he misses the main components of what makes a cloud a cloud.

I recently attended a presentation with Lydia Leong of Gartner who defined the core attributes of a cloud in a very succinct and clear way. According to Gartner, the key components of a cloud are the following:

  1. Service-based: this means that the user of the cloud solution need only worry about the service that is being purchased, not how it is delivered or how it is produced
  2. Scalable and Elastic: services scale as needed without intervention from the consumer of the services
  3. Shared: resources are pooled for maximal efficiency
  4. Metered: users pay for what they use
  5. Internet-based: delivered over the web using standard interfaces and APIs

So Larry is right that cloud services are delivered over the internet but he misses the point. What makes the cloud transformational is its ability to abstract out the activities that occupy much of CIOs’ time today, i.e., managing, provisioning and scaling infrastructure (perhaps Sun servers) or databases (mostly Oracle) and middleware (Oracle or IBM).

“They change a term and they think they’ve invented new technology. Let’s call that cloud, it sure beats innovation.”
Our assessment = 0% accurate for true cloud providers, 100% accurate for traditional vendors jumping on the bandwagon
Companies like Saleforce, Amazon and Google have invented fundamentally new technology. Whether it’s specialized hardware or massively scalable non-relational data structures or multi-tenant platforms with metadata-driven customization, these companies have introduced fundamentally new technologies and computing paradigms.

In addition to core technology innovations, each of these vendors delivers more innovation to their customers at a fundamentally different rate than traditional software companies. Salesforce delivers 4 major releases every year and Google introduces new features seemingly on a daily basis. Contrast this to the standard 3-5 year release cycles of the legacy vendors. In the past year alone, Salesforce introduced Force.com sites to power public websites, integrations with Facebook and Amazon and the Service cloud. Google introduced Calendar/task gadgets, video conferencing, Outlook synchronization, iPhone integrations and lots more. The best part of many of these innovations is that existing customers get new functionality instantly. No costly upgrades, no license re-negotiation, and no expensive user-training.

However, I have to agree with Larry if he’s talking about traditional vendors like his own company. Traditional on-premise vendors have jumped on the cloud bandwagon by introducing confusing notions like private clouds (i.e. datacenters) and relabeling hosted software as cloud software (see below).

“We offer our ERP system in the cloud”
Our assessment = 0% accurate
This was the boldest statement of the day and a perfect example of what Larry decried earlier in his diatribe.

There are three main reasons why hosted software is not cloud computing:
1) Single tenant: Hosting requires individual instances of software, middleware and databases for each customer. There may be some economies of scale at the HW level but that is it. Compare that to Salesforce CRM, where every customer is running on the same version of the software, middleware and database. This has massive implications for cost and scalability since a traditional vendor needs to maintain a new stack for each customer and all a true cloud vendor needs to do is maintain one stack.
2) Slow pace of innovation: Since hosted SW is traditional software with all the attendant complexities, the pace of innovation is a lot slower than in the cloud. Hosted SAP or Oracle will be upgraded every 3-5 years, not every quarter or sooner.
3) Customization: A big part of hosting providers’ economies of scale come from the ability to maintain standard environments. Since traditional software is not customized at the metadata-level, standardization implies limited customization for customers. This is quite different than applications like Salesforce CRM that can be customized to meet customer requirements at the metadata-layer without altering the foundations.

Summary
Treat these pronouncements about the cloud for what they are. Entertainment.

The true measure of the impact and relevance of cloud computing isn’t what anyone says, but what CIOs are actually doing. 90%+ of enterprises plan to maintain or increase their investments in cloud technologies. Salesforce is growing at 20%+ while Oracle and SAP contracted at double-digit rates. That’s the true bottom-line on cloud computing!

Previous Article
Learning from the Pioneers – Schumacher Group
Learning from the Pioneers – Schumacher Group

Today, we’ll be speaking to Doug Menefee. Doug, a Gulf War veteran, has been involved with technologysince ...

Next Article
Summary: Questions asked of the Enterprise SaaS Working Group
Summary: Questions asked of the Enterprise SaaS Working Group

Ryan Nichols Great kickoff today of the Enterprise SaaS Working Group, sponsored by Conformity. With Scott ...