In a previous tip on the challenges of performing disaster recovery across the WAN, I discussed how data replication and backup can be difficult to accomplish reliably because a large amount of information is sent across low bandwidth/high latency links.
WAN acceleration can be an essential tool for overcoming these challenges. More specifically, technologies such as data reduction, compression, and latency mitigation can improve disaster recovery efforts in the following ways:
Improve data transfer times
Data reduction is a new WAN acceleration technique that recognizes information and delivers it locally whenever possible. This can reduce file transfer times from several hours to a few minutes in environments that are highly repetitive, as is the case with replication/backup. In addition to data reduction, enhancements can be made to TCP, including selective acknowledgments and window scaling, which help to overcome the impact of latency across a WAN. As a result, large files such as VMware system images can be transferred 50 to 100 times faster when WAN acceleration is employed.
Maximize WAN efficiency
Data reduction can reduce WAN traffic by as much as 99%, improving overall WAN efficiency. By fingerprinting at the byte level, WAN acceleration can detect and eliminate repetitive patterns even when the backup/replication solution is performing similar functions at the block level. In addition, advanced compression techniques can be used to further reduce the amount of WAN bandwidth required for backup and replication. These techniques work in both directions of a WAN link, improving the backup/replication process while ensuring that the WAN can efficiently handle a restore if needed.
WAN acceleration can provide 10-100x reductions in WAN traffic when handling disaster recovery applications. That can make a 45 Mbps WAN link look 450 Mbps or higher.
Reduce packet loss and delivery errors
WAN acceleration can reduce the impact of both packet loss and jitter that occurs when router links are oversubscribed and drop or reorder packets. Adaptive Forward Error Correction (FEC), for example, can reduce effective packet loss by an order of magnitude by injecting a FEC packet for every "n" regular packet. This packet can be used to reconstitute a lost packet at the far end of a transmission, eliminating the need for packet transmission.
Increase geographic distances
By reducing the impact of latency, enterprises can extend the distances between data centers and disaster recovery locations, increasing operational flexibility.
Data reduction, compression and latency/loss mitigation are all essential for overcoming WAN performance limitations. This is particularly true when handling large volumes of data, as is the case with disaster recovery. WAN acceleration can be deployed in a variety of backup/replication environments, including data center to data center transfers and network backup across many remote sites. Look for solutions that work on both TCP and UDP traffic, because they are a perfect complement for all disaster recovery applications, regardless of configuration.
About the author:
Craig Stouffer is vice president of Worldwide Marketing at Silver Peak Systems. He has 17 years of industry experience in marketing, product management and business development. He previously held positions with Juniper Networks, Redline Networks (acquired by Juniper), Optranet (acquired by Extreme Networks), and Tut Systems.