Discussions of Cloud computing often revolve around strategy, business, economics, and technology, particularly in the higher levels of the Cloud stack, SaaS, PaaS, so you may wonder why I’m writing about Intel’s Day in the Cloud event when microprocessors are found deep within enterprise infrastructure components, and the Cloud computing industry focus seems to be largely on Software-as-a-Service (SaaS) applications.
To better understand the connection between semiconductors and Cloud computing, I invite you to read Silicon in the Cloud. The focus of that blog post is on the influence that the semiconductor industry has on the speed of development and capabilities for Cloud computing as a whole.
The net of the article is that the semiconductor industry influences over $1 Trillion in electronics and $5 Trillion in services or about 10% of global GDP, and if it were not for multi-core developments in microprocessors, we may not be discussing virtualization or cloud computing today.
To give you an example, semiconductor developments will drive increased Infrastructure-as-a-Service (IaaS) capabilities enabling features and functions at higher levels in the Cloud stack that rely on the underlying technology. Let’s take virtualization, for example.
- Increased capabilities in microprocessors that enable VMWare to manage thousands to tens of thousands of virtual machines increases infrastructure manageability while lowering costs for client service providers like Amazon.
At this point some of you may be calling me captain obvious, though I can’t tell you how many times I’ve been in discussions regarding infrastructure and heard industry colleagues say that the resulting decoupling of infrastructure from applications means that infrastructure no longer matters. To a degree, that’s true in the sense that applications are no longer tied to the underlying infrastructure technology the same way legacy IT was used to developing those applications, that’s the result of programming languages and application frameworks like Ruby on Rails and others.
At this point Cloud discussions often devolve into the merits of flexible and agile application development with Ruby, Python, Java, and the use of REST APIs, JSON, XML, and Hadoop, etc. I’ll let others more informed on these topics debate those issues.
I prefer a more holistic business and technology approach to Cloud computing since there are many issues to consider in both the business and technology domain.
- On the technology front, this increased capability to manage tens of thousands of VMs translates into ever greater economies of scale placing increasing pressure on Amazon Web Services’ hosting competitors.
While those same competitors are leveraging these same technologies, Amazon’s resources are more extensive allowing development at blazing speeds. (Side note: Can competitors catch Amazon?)
Smaller local and regional managed service providers (MSPs) will have to differentiate on other variables such as relying on their existing client relationships as trusted advisors, but as manageability increases a larger organization like Amazon will be able to reach further into SMB putting tremendous pressure on mid-sized and larger MSPs.
Additionally, Amazon enables new smaller MSP entrants to come into those regional markets at ever decreasing price points given their continuous ability to scale as a result of microprocessor and virtualization developments.
- Ultimately, a well-played technology advantage will translate into a business advantage, at least until adopted by the competition.
And that’s why in my next post I’ll be writing about Intel’s Day in the Cloud event because sometimes the infrastructure you’re running does matter.