
|
 |
|
General Information
|
Add To My Personal Library |
January 23, 2009
Vol.31 Issue 4 Page(s) 18 in print issue
|
Cloudy Days Ahead
A Look At The Different Layers Of Cloud Computing
|
Over the last few years, the buzz has been about virtualization. While much of the initial focus has been on individual servers and consolidation, the virtualization of the data center as a whole has given rise to the commoditization of applications through the elimination of tying software to the physical hardware. From this simple concept, new paradigms revolving around the cloud (aka the Web) are gaining in popularity throughout the enterprise “One of the reasons that I think the cloud metaphor is gaining such traction is because there are a number of different layers that can be expressed and tapped into. That has produced a lot of stimuli for people to think about elastic services that are applied outside of the normal methodologies and realm that they normally think about,” says Steve Oberlin, chief scientist with Cassatt (www.cassatt.com).
Cloud Flavors Software as a service is essentially the outsourcing of an application to a cloud provider where the application is hosted on an external service. “You don’t necessarily know or care where your application is running or how those systems are being managed. You are being charged for the use of that end application being provided to you,” says Oberlin. “There are many organizations choosing to actually move to cloud-based services for things that are traditionally in-house, such as email. This means someone will run Exchange out in the cloud rather than having their own IT administrators worrying about Exchange. Now, you can pay somebody on a per-mailbox basis to do that for you,” says Stephen Pao, vice president of product management for Barracuda Networks (888/268-4772; www.barracudanetworks.com). The next level in cloud computing is platform as a service, or PaaS, and refers to when the enterprise owns the application but is writing it to a particular platform with well-defined APIs that are hosted by a cloud service provider. Microsoft recently announced its Azure Services (microsoft.com/azure), based on the Microsoft .NET Framework that is poised to deliver interoperable services that can run on the Internet and within internal environments. Google is also offering a similar service, except its API is through Python, a common scripting language. “The enterprise writes their application, but the capacity that powers that service is provided by whoever is running the software stack all the way up to the presentation layer,” says Oberlin. There is also IaaS (infrastructure as a service), in which the enterprise provides the application and software stack based upon a particular infrastructure or server definition. The server is defined and supplied by a third-party supplier based upon the users’ needs. An example of this is Amazon Elastic Compute Cloud (aws.amazon.com/ec2), which allows users to only pay for the resources that are actually consumed, such as instance-hours or data transfer—essentially, renting servers over the Web. “Users can purchase as many servers as they want where the servers are defined by different scales depending on network connectivity, memory capacity, CPU, etc. It’s defined as a particular unit of server horsepower,” Oberlin says. “You just set up an account and pay for [the servers] with a credit card. The sky is the limit, and when you no longer need them, Amazon will take them back.” This extremely dynamic hosting service provides an elastic unit of service in the cloud and is defined as a particular class and nature of a server. Unlike signing up with an ISP or a colocation facility, enterprises are not required to sign a long-term contract or only receive a particular section of a dedicated server. “I can now say I need a hundred servers and then give them back three hours later. There are very few ISPs even approaching that magnitude of flexibility or ability to provision servers on demand,” says Oberlin. Moving beyond traditional IT offerings, content as a service, or CaaS, is also emerging from the cloud. Publishing and marketing on-demand have led to the need to pull information from multiple and disparate virtual locations in order to facilitate real-time business models driven by the management of Web content.
Challenges While many of these cloud mechanisms are being employed, there are still a number of challenges facing organizations wanting to move in that direction—namely security, service levels, and regulatory compliance. Pao has witnessed much confusion with regard to effectively securing applications and information in a cloud environment. He asks, “If your applications and data are in the cloud, does it make sense to have your security be on premise?” Organizations turning to cloud computing must think through all the implications of moving critical applications out of their physical realm, or they could face legal repercussions due to a lack of audit ability for stringent compliance regulations. Additionally, in order to adopt an external cloud supplier on any one of these levels, enterprises must first migrate their internal capacity to a cloud product. “If your applications and the platforms they’re written for don’t match a particular cloud vendor’s offering, you may not be able to move it into the Web. The same is true for infrastructure as a service,” warns Oberlin. As a result, the cloud is expected to be an asset for new applications and the elasticity aspect for giving access to capacity in many cases for unplanned demands. But there's always going to be a need for internal or private clouds, which is leading to the next revelation—internal cloud computing. Oberlin explains, “Instead of resource stovepipes that are isolated and hard-installed with the application, you adopt a cloud-like model—turning your own data center into a cloud that your own enterprise can treat as an elastic resource.” by Sandra Kay Miller
Types Of Cloud Computing Software as a service: The end application is purchased on an as-needed basis. Platform as a service: The application is written for a particular hosted platform. Infrastructure as a service: Server resources are available on an as-needed basis. |
What Works Well In The Cloud Here’s an idea of how organizations are using cloud computing today. • Business applications (email, CRM) • Web servers • Downloads • Storage • Backup/disaster recovery |
|
|