Why Open Source and Operations Matter in Cloud Computing

As the IT industry shifts from a product base economy to a service-based economy, operational competency is a competitive weapon. Contrast this with the past where companies could rely on closed-APIs, vendor lock in or the reliance on vast resources to build business and keep out the competition. Today, anyone with a good idea can connect to a cloud provider and build a software business over-night –- without massive investment dollars. Instead of forcing people to pay for a CD with your software on it, you deliver a service. In that type of environment where service is king, operational efficiency is crucial. It’s the company with the best execution and operational excellence that prospers. Yes, it’s leveled the playing field, yet ironically the cloud providers themselves are the best examples of operational excellence being the competitive advantage of the 21st century.

There are a few companies who can affect these economies of scale and create these cloud offerings: Google, Amazon, and now IBM. All of these cloud offerings run on Linux. (Microsoft will likely enter the fray with a Windows-based offering at some point.) These large giants who have built their own businesses via their operational excellence, now have the ability to drive down the cost of computing per CPU to rates that no one can compete with on their own. As InformationWeek writes: “IBM said it believes customers can cut IT labor costs by 50% and reduce software defects by 30% by moving development to the cloud.

The problem with internal development and test environments, IBM said, is that they consume as much as 50% of an organization’s entire IT infrastructure but typically remain idle 90% of the time.”

In typical IBM fashion they have focused on those workloads that make the most sense for Cloud computing. Their offering will allow customers to focus their own operations on production environments, while reducing costs in development and test services.

Why Linux?

Linux is the operating system of the cloud. Why is every cloud provider using Linux?

• Linux can be optimized for powerful parallelized computing to run these types of environments efficiently. IBM is using KVM, built into every Linux kernel, to power their offerings and partnering with Red Hat, a company based entirely on open source and known for its technical skills and high levels of service.

• Linux has tremendous power management capabilities. This is due in part to the focus on enterprise Linux by companies such as IBM, Red Hat and Novell to bring the cost of running a data center down. But Linux also benefits from technical innovation by mobile/embedded developers who are using Linux in those devices and need advanced power management features.

• Linux because it is open and not optimized for a specific architecture can run on a multitude of hardware options, bringing down the price for the vendors building these massive data farms. Because Linux is open, IBM has optimized Linux for its mainframe computers, giving them an advantage operationally.

• It’s pricing model. You can’t build a data farm and charge $.15/per CPU hour if you have to write a check to Redmond or anyone else for every server. Microsoft may be able to run a cloud business with MSFT products, but anyone else who actually has to pay for them would not.

• Linux is the development platform of choice in today’s world. IBM couldn’t offer a Linux-based development and test Cloud service if companies weren’t developing on Linux. Linux has steadily gained momentum as the development platform of choice, largely due to the points mentioned above that make it a great candidate for cloud computing platforms.

• Ownership. I experienced this in the late 90’s as one of the founders of a “cloud services” company called Corio, which offered hosted enterprise applications (we called cloud “application service providers” then). We were required to disclose a risk in our public offering S-1 filing that stated “We depend on software vendors to supply us with the software necessary to provide our services, and the loss of access to this software or any decline or obsolescence in its functionality could cause our customers’ businesses to suffer, which, in turn, could harm our revenues and increase our costs.” If we had run our business on open source software we would have owned our own software and this risk would not have existed. This may be the single biggest advantage to Linux in the cloud. Ask yourself if Google could be the company they are today if their search engine was built on .NET servers.

The nascent history of our service-based economy is littered with companies who have failed because of a lack of operational excellence. One example: Friendster. Back in 2003, Friendster had all the momentum. Before Facebook or Twitter or even MySpace, Friendster had amassed the first-mover advantage which is usually so important. Unfortunately, the company didn’t scale and delivered innumerable uptime problems. (I remember since I was an early subscriber.) They let poor development and operational issues sink their site, providing a bad customer experience. Users flocked to Facebook, Twitter, et al, and now Friendster is a marginal regional player at best. There were no cloud offerings at the time to help Friendster scale. Perhaps if they could have outsourced their development and test operations to IBM and focused their efforts on production, the Linux Foundation would be touting its Friendster page instead of its Twitter and Facebook accounts (nearly 25,000 members by the way.)

I have a feeling this is just one of a series of cloud announcements by IBM. We are pleased to see another Linux-based cloud offering now available to create innovative service-based companies in the future.