Who uses continuous availability, and why
Continuous availability architectures and processes are used by forward-thinking organizations that seek to ensure exceptionally high levels of service availability for their mission-critical and business-critical applications. Two factors are driving this migration to continuous availability solutions. The first driver is the escalating demand for uninterrupted availability of key business applications, as consumers and markets have come to expect that services will be accessible anytime, from anywhere, without exception. In this context, the costs of even brief outages can be severe.
The second driver is the shortcomings of the conventional approach to the availability challenge, which is to pair separate and distinct systems and processes for high availability (within a single site) and disaster recovery (across primary and backup sites). The flaws of this approach include:
Excessive complexity and cost, as administrators deploy and manage disparate systems while standby resources sit idle unless needed.
The expectation, despite the expenditure and management effort, is that services will nevertheless suffer some costly period of downtime in the event of a disaster.
How continuous availability works
Continuous availability is made possible by technological advances that make it feasible and economical to build and maintain multi-site active/active architectures for transaction-intensive business applications. Rather than having costly resources idling in passive mode, continuous availability topologies efficiently “stretch” active assets across two or more sites.
The main elements of a continuous availability architecture are:
Two (or more) data centers deployed in active/active mode, each with complete infrastructures that have no single point of failure. The sites need to be far enough apart to not both be impacted by a single disaster, but close enough to allow for synchronous data replication across sites. Depending on regional location, between 25 and 60 miles apart is the “sweet spot”.
An advanced distributed data virtualization layer such as EMC’s VPLEX Metro to synchronously maintain data coherence across sites.
Business applications architected to run in continuous availability mode, using an intelligent clustering DB solution like Oracle RAC.
A transaction distribution mechanism to balance loads across the active/active sites during normal operations, and to route all traffic to the good site in the event that disaster takes a site down.
Benefits of continuous availability
Automating the data archiving process and using purpose-built archive systems make production systems run better, use less resources, and reduce overall storage costs. Production performance is unaffected by information growth. Backup and recovery runs faster, disaster recovery is less costly, and systems are easier to manage. Data moved into archives is stored at much lower cost.