It’s no secret that enterprise applications are increasingly being operated in ‘cloudy’ environments. When systems are moved to the cloud, they often are moved partially or relocated in stages that are instep with the evolution of other IT systems. Even when traditional back-end enterprise systems such as ERP, data warehouse and similar systems remain entirely in-place, they’re increasingly being augmented or extended with vendor software and services that reside in the cloud. For example, many organizations have already integrated Salesforce.com into their back-end enterprise systems, and these back-end systems often reside on premise or in a colocation environment. Thus, Hybrid IT is already the current reality for many IT organizations and is on the road to becoming ubiquitous. Optimizing Hybrid IT architectures, from an availability perspective, is the central question considered here. If Hybrid IT is the new normal, then what are the implications for how we design and manage systems that run across multiple data centers in different locations? In a new whitepaper titled the ‘Seven Secrets to High Availability in the Cloud’, lessons are borrowed from distributed computing and applied to Hybrid IT scenarios which reveal opportunities for improving availability, despite growing complexity.

The ‘Seven Secrets to High Availability in the Cloud’ whitepaper calls out the following opportunities to improve the availability of your hybrid applications.

  1. Use a managed DNS service
  2. Assess failure risks
  3. Calculate availability
  4. Balance loads
  5. Make use of horizontal auto scaling
  6. Make use of vertical auto scaling
  7. Manage storage for HA

There’s no question that evaluating application reliability and availability becomes a more complex exercise in a Hybrid IT environment. However, by modeling the resultant hybrid system, using techniques that are well known in the domain of distributed computing– overall availability can be improved despite greater overall system complexity. Optimizing hybrid architectures to gain availability appears to be a new muscle that many organizations will have to develop and are likely to exercise often.