Management
The Architecture Behind It All
It would be easy to ignore the technologies behind cloud services, but it also would be a mistake. Business technology pros must ensure that cloud services integrate with their enterprise infrastructures. That requires an architecture that spans both.
The components of cloud computing are the same as those in data centers: programming languages, operating systems, databases, Web servers, protocols, APIs. The task is to identify cloud services that are a good fit with your internal systems, applications, and expertise. A comparison of Amazon's Elastic Compute Cloud, Google App Engine, and Windows Azure services shows how that might work.
Amazon's EC2 lets customers pick from a software smorgasbord: Windows Server, OpenSolaris, and seven Linux flavors; the MySQL, SQL Server, and Oracle 11g databases; and the Java, JBoss, and Ruby on Rails development environments.
Google's forte is simplicity. App Engine lets users tap into Google's homegrown database and other infrastructure software, and APIs provide access to caching, imaging, mail, and other application services. Python is the only programming language supported, though Google intends to add support for others in the future.
Windows Azure and Azure Services Platform are cut from the same cloth as Microsoft's on-premises enterprise line. Azure comprises hosted versions of SQL Server, SharePoint, Dynamics CRM, and .Net Services, and it's developed in Visual Studio and the .Net Framework. Microsoft says Azure will support open protocols (HTTP, REST, SOAP, XML) and non-Microsoft languages (Eclipse, Ruby, PHP, Python).
For IT pros who need to sketch out a cloud architecture, much of the granular information needed is available on service providers' sites. Amazon has a white paper on cloud architectures that's worth a read for anyone trying to come up to speed.
Your blueprint should take into account the possibility of cloud services from multiple vendors, so think about how you would accomplish interoperability and application integration. Stuart Charlton, senior software architect of cloud computing startup Elastra, recommends REST and the Atom Syndication Format as underlying specifications in a global cloud architecture. Standards for federated identity management also are key, he says.
Dennis Quan, IBM Software's director of development for autonomic computing, says service-oriented architectures already make it possible to connect cloud services in "standards-complaint ways." The next trick will be to transplant services from one cloud to another. The specs to do that, Quan says, are still in their infancy.
-- John Foley
Bob Friday, Chief AI Officer for Juniper Networks, explains how the advanced technology is transforming operations.
Contact center leaders from 8x8, Awaken Intelligence, and 360insight discuss the importance of agent experience.
AI may force enterprises to rewire parts of their data centers so they are fully optimized to run such workloads. The question is do you use Ethernet or InfiniBand?
