Disaster Recovery

If you want to optimize your disaster recovery capabilities…You need Ipanema!

You are running the network of a large international company. The Enterprise recently has decided to build disaster recovery capabilities. Primary datacenters are backed by secondary datacenters hosting a replica of the business critical applications. The secondary datacenter must be ready to deliver to each user in the Enterprise the content they need to maintain productivity in the event of a disaster.

Disaster recovery architectures are directly impacting the WAN design and capabilities with mainly two issues. First the sudden change in terms of the traffic matrix over the WAN caused by a switch to the backup datacenters. Second the increased strain put on the WAN to serve the purpose of datacenter to datacenter replication, which is mandatory for a smooth transition between primary and backup datacenters.

The redirection of thousands of users from a primary location to a secondary location represents a huge shift in terms of the WAN traffic mix that traditional QoS mechanisms such as MPLS Class of Service cannot deal with. The company datacenters are typically not limited to one and as a result the flows across the WAN are meshed. Meshed flows are difficult to handle in normal situations and in most case impossible to manage when such massive changes occur, leading to unpredictable behavior of business critical flows.

Datacenter replication is also pushing the network to its limit. The secondary datacenter must keep a fresh copy of the application content and state. To do so, high speed links are used between the distant datacenters. Those are referred as “long fat pipes” as they exhibit an extreme Bandwidth and Delay Product characteristics. Over those long fat pipes, different backup applications are used (EMC, Network Appliance, Legato, Veritas, Tivoli, Microsoft, CommVault, DoubleTake, CA …). All rely on the TCP protocol, which is not optimized for long fat pipes … And most of them do not fully optimize their bandwidth requirements with application level compression. As a result datacenter replications are slow and business continuity is jeopardized.

The Ipanema Solution is the only solution on the market tackling both disaster recovery WAN related issues. The dynamic and global nature of its Optimization functions make it capable to fully controlling meshed flows. Its ability to reassess the optimal traffic management policies at each second and its layer 7 classification capabilities allow an immediate and automated reconfiguration of QoS across the whole network to cope with a sudden change in the traffic matrix. As a result the disaster recovery capabilities are not impaired at all by the WAN service. The complete Acceleration features of the System mitigate both bandwidth and TCP limitations of backup applications for a much faster replication. The TCP Acceleration allows to make full use of long fat pipes and the Multi-Level Redundancy Elimination dramatically compress the backup flows leading to a massive “virtual” bandwidth being available to perform the backup.
 

Ipanema comprehensive solution with its unique ability to guarantee WAN application performance whatever happens is the ultimate companion to your disaster recovery compliant datacenter design.