WAN Optimization Manager™ (WOM) uses symmetric data deduplication to reduce the amount of bandwidth consumed across a WAN link for repeated data transfers. This feature is available only with a WOM® license.
With data deduplication, the system performs pattern matching on the transmitted WAN data, rather than caching. If any part of the transmitted data has already been sent, BIG-IP® WOM replaces the previously transmitted data with references. As data flows through the pair, each WOM records the byte patterns and builds a synchronized dictionary. If an identical pattern of bytes traverses the WAN more than once, the BIG-IP WOM closest to the sender replaces the byte pattern with a reference to it, compressing the data. When the reference reaches the other side of the WAN, the remote BIG-IP WOM replaces the reference with the data, restoring the data to its original format.
Symmetric data deduplication (SDD) offers two versions, called codecs. SDD v3 is appropriate for most WOM installations, particularly in large networks, such as hub and spoke, or mesh deployments. SSD v2 is an alternative for installations with fewer than eight high-speed links, such as for data replication between data centers.
For deduplication to occur, the same codec must be selected on both iSession endpoints. If the selected codecs do not match, deduplication does not occur, although other WOM features, such as compression, still take place.
|Disk||Specifies that WOM uses the disk, in addition to memory, for storing
information used for optimization.
Note: If you enable data storage on the disk, you must restart the datastor service from the command line using the command sequence bigstart restart datastor for the change to take effect.
|Memory||Specifies that WOM uses only memory for storing information used for
Note: This setting can provide benefits for higher speed links.
|SDD v3||Supports a high spoke count, such as for connecting remote sites and for mesh topologies.|
|SSD v2||Supports a topology with fewer than eight spokes, such as replicating data between data centers.|