Companies are Still not Extracting the True Promise of Cloud
Date: Friday , April 08, 2011
Cloud computing will become as ubiquitous in next decade as renting a server or a space on a rack became in the last decade. Perhaps even more as private clouds start gaining acceptance and Cloud computing will not simply remain something ‘out there’ but inside the company offices. Although large companies will be early adopters, even mid-sized companies will soon be using mature open source solutions like Eucalyptus very soon for creating private clouds.
How we came to where we are?
As most of us know, Cloud Computing is the same promise that buzz words of yesteryears called Utility Computing and Grid Computing wanted to provide. The names themselves clearly market the value: computing on tap like the other utilities: Gas and Electricity. The technology was pretty much perfected independently by multiple providers like Amazon and Google for their growing internal infrastructure needs. Creating farms of redundant low cost commodity linux servers backed with intelligent virtualization and redundancy software layer on top allowed them to scale cheaply for huge growth these players experienced. Then Amazon changed the game by exposing its infrastructure as a cloud service to other providers who did not want to invest in creating such capabilities themselves. Amazon also labeled its service: Elastic Compute Cloud or EC2. Thus the name Cloud stuck.
The cloud fundamentally allows an elastic utilization with the underlying platform supporting hourly pricing. However, even with the hourly pricing schemes most companies are still not using the elasticity of the platform to the extent that say electricity consumption is elastic where Air Conditioning maybe switched on during day and only fans maybe good enough at night. Say, you have an India specific news portal that you host in the Singapore center of Amazon (lowest latency from India). Typical traffic patterns show that news portal traffic peaks between 4pm to 6pm when employees in office want a break from work and check news. Ideally we would like server capacity to be maximum between these 2 hours, medium during rest of the day and minimal during night. This will of course be backed with ability to respond dynamically to unforeseen sharing of a news link on Digg.com. One would think that this is what everyone ought to be doing with the promise of hourly elasticity. The reality is that tooling has not yet caught up to make these scenarios trivial. Result: we are still not leveraging the full savings that the cloud enables and just scratching the surface.
Companies like Neev Tech(www.neevtech.com) are breaking new ground by creating innovative tooling and enabling companies to experience the full savings that the elastic platform promises. The graph in the article shows what is the potential of using the computing power and storage truly like an utility e.g. electricity.
An Elastic Cloud Strategy thus gives companies an opportunity to massively reduce their infrastructure related expenditure while also maximizing the savings that can be made on the cloud. The space is currently open for a lot of vendors to provide the next level of innovation on top of the cloud infrastructure and extract the true promise of elasticity. The cloud providers are themselves improving the tooling available out of the box but the vast variety of enterprise architectures will need professional services help to effectively utilize these cutting edge platform features. Apart from elasticity, monitoring is another significant area especially when we include performance monitoring and not just uptime monitoring in the purview. This is especially critical for enterprise applications where server infrastructure of mobile and sales applications can be very performance sensitive.
The author is Co-Founder & CEO, Neev Information Technologies