In the Clouds
Posted: February 26th, 2010 | Author: Julio Hernandez-Miyares | Filed under: Cloud Computing |
At least in the technical realm, the noise around cloud computing has become a veritable roar with all of the usual hype associated with the promise of alleviating some of the more difficult aspects of operationalizing the assorted services that constitute the web as we know it. In many respects the hype is legitimate and is making technologists’ potential to concentrate on the core feature set and leave the operations to someone else that much greater.
It the context of this post I will differentiate between Cloud Applications like Flickr or Delicious and countless others that provide product functionality directly to a consumer requiring precious little if any technical acumen and Cloud computing services like Amazon Web Services that provide a set of fundamental building blocks to build products.
Before I move on, a Cloud application I have recently started using that leverages Cloud services namely Amazon’s is Backupify. As the name implies, it backups various well known applications such as Flick, Delicious, wordpress blogs etc and the backups are persisted via Amazon’s Web Services. So here is an Cloud application that uses Cloud services to flesh out it’s features. The power of this model of crafting products is enormous. The time to market is considerably reduced as the purveyor of services can quickly spin up a set of functionality without much regard to establishing a data center/rack space and all that is involved in operationalizing something of this sort. Also, the vendor of the service is able to manage costs and attendant risk during the hectic startup period. They generally only have to pay for as much service as they consume which is governed by the adoption rate of new customers. This makes the operational costs almost entirely variable instead of fixed which provides the necessary risk mitigation especially for new startup like businesses. If the adoption rate is low and the service does not win out in the marketplace, there is no need to sell equipment at fire side sale prices to wind down an unsuccessful business. Conversely, growth can be managed in a more balanced manor by only tapping into the services needed. Also, for those particular instances where one has to deal with spikes in usage , the need to have an operational footprint large enough to support the anticipated spike ends up revolving around tapping into extra services from the Cloud to support the spike and when the spike is over, the horsepower is made available to whomever else out in the cloud when need the capacity.
This makes history of the scenario where one would need to manage fixed operational infrastructure sufficient to deal with the highest volume of usage but where that infrastructure is often left under-utilized save for the few instance of spikes.

Leave a Reply