The difference between new and not-so-new technology has a way of revealing what is elastic and dynamic compared to what is rigid and static. It’s not a measure of which technology is considered good or bad. It simply represents the progression from client/server technology to the Internet-scale, data-driven services that are gaining such momentum.
Using antonyms helps better correlate what is considered a cloud service and what is not, as well as the relative relationship between an online service like Google Docs as compared to a Microsoft Word document. The differences can help understand the new way IT services are delivered as compared to older methods.
Randy Bias, founder of Cloudscaling, did a keynote at Interop’s Enterprise Summit two years ago and argued that elasticity is a side effect of cloud computing. The infrastructure of cloud/web scale operations is fundamentally different from mainframes and client/server technologies.
The big Internet companies have had to create an infrastructure that could scale and be highly efficient and fast. The result: new ways to think of how we manage data.
Mark Thiele is executive vice president at Switch, which runs “SuperNap,” one of the most sophisticated data centers in the world. He said this to me today on Twitter:
The reality of elasticity is that it has metrics like any service. The speed/scale of elasticity should match need & ability to pay.
Elasticity is a side effect of Internet-scale businesses. But with the introduction of the smartphone, the notion of elasticity becomes more abstract. A smartphone is a node. It is a dynamic object existing on an elastic infrastructure. But it is also inelastic, often running enterprise software managed by IT.
As another example, Google Docs runs on an elastic infrastructure. Microsoft Word running off the client server is not.
Mark Thiele questioned this view of elasticity when he asked me this question in a private exchange on Twitter:
Definition of elastic comes in to play as well. If I can download a copy of MS Office & start using in 30 mins, is that elasticity?
I responded that the antonyms may not always be relative. They may still have an opposite meaning but exist in what Wikipedia refers to as a common continuous spectrum. A downloaded copy of Microsoft Word is managed through the elasticity of the more modern infrastructure.
Pradeep Prabhu is the co-founder at CloudMunch, an end-to-end platform for managing development and application lifecycle management development. Pradhu views the traditional developer software stack as not elastic.
In most software development environments, there is no elastic provisioning of systems for Continuous Integration (CI), Automated Testing and Automated Deployment. Systems are configured & activities executed manually and are not parallelized, not scaled dynamically and are not deployment agnostic. Also missing is an automated software delivery pipeline that is created dynamically when code is pushed by a developer.
Cloud services connecting with other cloud services is elastic, Prabhu said.
“When code gets developed by a developer — they get an elastic infrastructure and tooling that will take that code and build it, test it and deploy into the cloud,” he said. “It is completely elastic.”
Elasticity has a meaning relevant to the way IT services are delivered. Understanding its oppposite meaning can help clarify the difference and similarities between Internet-scale architecture and its client/server counterpart.