PRA: To ensure the orderly restart of applications in case of failure or disaster
For companies that do not have the means or the need for a PCA, the PRA is the solution to ensure an orderly restart and as fast as possible to the company’s IT infrastructure in the event of an incident. This restart is performed generally on a backup site, the company’s ownership or provided by a third-party provider. The PRA defines the architectures, the means and procedures necessary to implement to ensure the protection of applications it covers. Its objective is to minimize the impact of a disaster on the company’s business. There are several restart modes: warm boot is based on a synchronous or asynchronous data copy of the main site. This is to rely on the last known consistent state of data as a basis for servers positioned on the backup site. Data replication (which can be ensured by backup and replication technologies such as Veeam Backup and Recovery, or continuous data replication technologies such as EMC RecoverPoint, AppAssure Dell, IBM Tivoli Storage Manager, CA ArcServe Replication or at CA DoubleTake at Vision Solutions) provides a quick restart standby servers in a state as close as possible to that which preceded the disaster. The RTO (Recovery Time Objective – time to put the application into production – is therefore minimal and RPO (Recovery Point Objective – the time between the last consistent state of the data and the damage) reduced to a minimum, often within minutes .
The situation is somewhat different in case of cold standby. This also concerns many companies do not have the financial and / or technical support for a PCA or a warm restart. In this case, the disaster restart uses the last backups made by the company. These backups can in the best case be replicas from a backup system deduplicated disk-based Data Domain as a bay or in the worst scenario, a simple tape backup.
In case of disaster, the company has to activate its backup site, restore scratch its data from their backup media (disk or tape) and reactivate its applications. This is the most economical solution to develop a PRA but it has a price in terms of RTO and RPO. OTN is the minimum recovery time data and commissioning of servers, to complex environments can mean several hours or even days. The good news is that the trivialization of disk backup solutions like Data Domain has greatly reduced the RTO (from 5:00 p.m. to 2:00 on average according to an IDC study, 2012).
The RPO is dependent on the frequency of backups. In the worst case, it can reach one or more days (particularly for applications with backup windows are long and are not saved once a day or less). Again deduplicated backup discs has improved the situation by reducing backup windows (11 hours with a library of on average 3 hours with tape backup systems deduplicated disk as EMC Data Domain, Quantum DXi or HP D2D).
Note that if the cold start was the rule for many companies, especially SMEs there are still 5 years, the spread of virtualization and networked storage has made available a warm start to a growing number of companies. The PCA is still not necessarily accessible to everyone – even if it has become commonplace for some applications – but it is now accessible to many medium-sized SMEs. The progress made by applications (for example with the integration of failover in most databases), for storage arrays, backup to disk solutions and virtualization technologies should make it accessible to the largest number Over the next few years. And the advent of CAC and PRA cloud services should also help further democratize a little more these technologies