By Peter Gilis
I bet you hear or read at least one message around cloud computing a day. Why should you read this article? What does it bring to your table?
Let me explain. I’ve been working for HP for more than 30 years, and spent most of my time in consulting organizations on optimizing infrastructure, a term I often use as it covers all aspects to reduce the costs of IT: consolidation, virtualization, automation of all IT resources—servers, storage, network—but also people and process aspects. And cloud is not something new in IT optimization; it’s an evolution. Standardizing, sharing and automating IT resources and process have always been there, but we had to wait until the Amazon’s and Google’s of the world to understand that IT rationalization could reach these extremes. And you don’t need to go to these public cloud providers to get the same benefits; you can realize them within you own IT by building you own private cloud.
First you have to find the right business case. Line of business developers are bypassing IT and putting development of applications—and sometimes confidential test data—in public cloud environments. Offering a similar service within your own IT will avoid this Shadow IT and will allow you to build the foundation for a more complete private cloud services offering. On top of that it often reduces costs directly and on the long term.
What I see when discussing these topics with my customers is that most developers have their own development server, and often each development server is different: server model, OS version and patch level, middleware version and patch levels, etc. This creates a lot of integration problems when all modules are being combined, requiring lots of testing and rewriting. On top of that the variation of OS and middleware versions will require the same variation in your production environment, resulting in higher costs of operations, due to the lack of standardization. Also, test servers are often not used; 80% of the year these servers are “sleeping”, waiting for new modules to be tested. Often these servers are oversized, just to make sure that consolidated and stress tests can be executed when needed. And are these servers still running the right OS and middleware versions, and the correct (security) patch level? Probably not.
Standardizing development environments and consolidating test environments into a shared pool of resources is often an easy case to start creating a private cloud environment. And can it be done in 30 days? Yes.
CloudSystem Matrix and CloudStart integration services deliver you an integrated and open cloud platform, up and running in less than 30 days after hardware and software installation. In this case of standardizing and optimizing development and test environments, your development users can request a compute service from a number of pre-defined choices, have the requested services provided immediately from the portal, scale and cancel the service, use the service without worrying about security, patching, backup and monitoring, and receive a regular report on consumption and chargeback.
CloudStart integration services include a series of workshops to shape and define the compute services choices (up to 4 templates in the services catalog) and the overall CloudSystem configuration and customization options. The agreed service templates will be implemented and automated, and proper hand-over is included to make sure IT personnel can manage, maintain, and extend the services offered to your development organization. And once you’ve proven to be a private cloud service provider for compute services, more complex—though still standard—services can be offered, e.g. shared database services, shared web services, shared application server services.