Friday, May 11, 2012

The Cloud Service Offerings and Deployment Models

Cloud computing has been an attractive proposition both for the CFO and the CTO of an enterprise primarily due its ease of usage. This has been achieved by large data center service vendors or now better known as cloud service vendors again primarily due to their scale of operations. Google, Amazon, Microsoft, and a few others have been the key players apart from open source Hadoop built around the Apache ecosystem. The cloud
service offerings from these vendors can broadly be classified into three major streams: the Infrastructure as a Service (IaaS), the Platform as a Service (PaaS), and the Software as a Service (SaaS). While IT managers and system administrators preferred IaaS as offered by Amazon for many of their virtualized IT needs, the programmers preferred PaaS offerings like Google AppEngine (Java/Python programming) or Microsoft Azure (.Net programming). Users of large-scale enterprise software invariably found that if they had been using the cloud, it was because their usage of the specific software package was available as a service—it was, in essence, a SaaS offering. Salesforce.com was an exemplary SaaS offering on the Internet.

From a technology viewpoint, as of today, the IaaS type of cloud offerings have been the most successful and widespread in usage. However, the potential of PaaS has been high: All new cloud-oriented application development initiatives are based on the PaaS model. The significant impact of enterprises leveraging IaaS and PaaS has been in the form of services whose usage is representative of SaaS on the Cloud. Be it search (Google/Yahoo/Bing, etc.) or email (Gmail/Yahoomail/Hotmail, etc.) or social networking (Facebook/Twitter/Orkut, etc.), most users are unaware that much of their on-line activities has been supported in one form or the other by the cloud.

The cloud application deployment and consumption was modeled at three levels: the public cloud offerings from cloud vendors; the private cloud initiatives within large enterprises; and the hybrid cloud initiatives that leverage both the public cloud and the private cloud or managed services data centers. The IaaS_oriented services offered abstracted (or virtualized and scalable) hardware—like compute power or storage or bandwidth. For example, as seen from its pricing tariffs webpage for 2009, Amazon5 offered six levels of abstracted elastic cloud compute (EC2) server power: the “small-instance,” “large-instance,” “extra-large instance,” “high-cpu instance,” “high-cpu medium instance,” or “high-cpu extra-large instance.” Each of these are accompanied
by appropriate RAM, storage, performance guarantees, and bandwidth support. The PaaS offerings are focused on supporting programming platforms whose runtime implicitly use’s cloud services offered by their respective vendors. As of today, these highly vendor-locked PaaS technologies have been leveraged to develop new applications by many startups. Compared to IaaS offerings, applications riding on PaaS deliver better performance due to the intrinsic cloud support for the programming platform. The SaaS on Cloud offerings are focused on supporting large software package usage leveraging cloud benefits. Most users of these packages are invariably ignorant of the underlying cloud support—in fact most, if not all, do not care. Indeed, a significant degree of the features of the software package invariably reflect the support of the cloud computing platform under the hood. For example, in gmail, users hardly bother about either the storage space taken up or whether an email needs to be deleted or its storage location. Invariably these reflect the cloud underneath, where storage (most do not know on which system it is) is easily scalable or for that matter where it is stored or located.

Source of Information : Wiley - Cloud Computing Principles and Paradigms 2011 

1 comment:

Blogger said...
This comment has been removed by a blog administrator.