Private Cloud Cuts Costs, Speeds Provisioning

As one of the fastest growing IT services companies in India, L&T Infotech, regularly undertook software testing and development projects for clients. As each client required a dedicated hosting infrastructure for itself, it led to under utilization of project-specific computing infrastructure.

As the number of projects started increasing, L&T Infotech found itself staring at one big challenge--on one hand, the need for servers were increasing everyday, while on the other hand, the average server utilization was as low as 20%. The Infrastructure Management Services (IMS) and IT Infrastructure division of L&T Infotech recognized this as a classic server sprawl and decided to build a private cloud christened CloudX.

L&T Infotech started the journey towards the private cloud by taking steps to consolidate and virtualize the servers. For the first phase, the firm decided to have only development and testing servers on the cloud. Production servers were only virtualized and were not part of the first phase. As L&T Infotech had a good blend of different hypervisors for different class of computing needs such as open source-based hypervisors (Xen & KVM) for development, testing and POC servers and commercial hypervisors (VMWare and HyperV) for production servers, the firm felt a need to have a private cloud solution that would be agnostic to underlying virtualization.

Since data centers were spread across various locations, the firm wanted to have a private cloud system that was geographically aware. "To save WAN bandwidth, a need was felt to have images provisioned locally. Hence we built a Geo-smart or location-aware cloud, so that users could provision images from the location they have requested it from," explains Abhay Chitnis, vice president and head of technology, L&T Infotech.

Post deployment, the private cloud has transformed the heterogeneous physical infrastructure into an infrastructure that is optimized for performance and cost. Business units have the ability to administer fixed amount of computing units, in addition to the capability to monitor the usage of provisioned units and use of licenses. The private cloud also has a metering and chargeback model in place. Users are charged based on the type of server, time-based usage of server and the software licenses used.

Built in governance policies for IT licenses, protection policies, role-based approver's delegation and integration with active directory ensure that the IT infrastructure can be focused on meeting service levels by taking a more strategic approach. The private cloud also has a self-service cloud management portal showcasing a catalog of services.

Read the rest of this article on InformationWeek India.

Data centers face increased resource demands and flat budgets. In this report, we show you steps you can take today to squeeze more from what you have, and also provide guidance on building a next-generation data center. Download it now.