The Business Situation
The requirements for highly available IT infrastructure in Datacenters are at an all-time unprecedented high mark. More and more applications are customer facing and utilized around the globe and around the clock. The key applications driving this requirement are primarily for online transaction processing of orders and payments. Additionally, with use of centralized corporate infrastructure to support the needs of global companies, the need to be available on a 24x7 basis is critical. When key applications are down, the value of the company does not exist for its customers.
The IT Challenge
Business needs that require applications with near 100% uptime requirements present unique challenges for the IT manager. Traditional tasks performed during regularly scheduled "maintenance windows" must now be performed without bringing down the systems. Not only do data back ups need to be done live, but routine tasks like system patching and application upgrades also need to be done live. Standard system and application management routines need to be modified to ensure that all critical systems are performing to the established service levels. The ability to monitor and automatically
restart application services is a required capability. In the case of a localized
power failure or other infrastructure outage, critical applications need to be moved temporarily to an alternate location. Once service is restored to the primary datacenter, the application delivery needs to be moved back to the primary location.
In the event that server virtualization is employed, this availability problem becomes more challenging. In a virtualized server environment, not only does the IT staff need to monitor and manage the application server, they need to watch the host environment as well. The virtualized network supporting the guest servers also presents opportunities for performance and response degradation.
Finally, in addition to server availability, all of the other IT resources need to be redundant, such as storage management, networking, and security, as well as the administration, support and management of the datacenter.
The dcVAST Solution
The first step to be considered as it relates to providing highly available infrastructure is to determine what the application requirements are. State-less applications just need component redundancy built into the design. Web servers and most application servers are examples of this kind of environment. State-full applications require that the application fail over and continue operation on an alternative server.
The next step in determining the proper solution to availability is determining the reasons that applications fail. Hardware failures account for a small percent of failures, as does external forces such and weather and power. Some system failures are also caused by software issues such as incompatibility between the software revisions and the hardware or operating system software. Finally, the largest reason for system failure is human intervention. These causes are typically someone making a change to software or performing maintenance activities and not following established processes.
The development of a proper availability solution is a combination of redundant architecture, clustering, and a good set of operational and support procedures.
dcVAST engineers will work with the customers to understand their application environment and the business needs of that environment. In this realm, one solution certainly does not fit all needs. The unique needs of each customer must be understood and a tailored program implemented to meet these needs. dcVAST utilizes best of breed products, and high availability best practices, assembled in an optimum configuration to meet your application availability requirements. To accomplish these goals, dcVAST uses a proven process five step process consisting of:
- Administration and Management