Technology |
Benefit |
How it Works |
Windows Scaling |
Link bandwidth is fully utilized. |
TCP connection window is adjusted to suit the Bandwidth Delay Product of the link. |
Selective ACK |
Better handling of packet loss & retransmissions. Deals with loss on the LAN. |
Notifies the host of what packets have been received subsequent to packet loss. Avoids retransmission of successfully received data. |
Packet Aggregation |
Better performance by reducing acks from lots of small packets. |
Small packets are aggregated into large packets before being sent on the WAN. This is especially important when data is compressed and de-duplicated. |
Congestion Control |
More suitable congestion control for a variety of environments |
Ability to select a congestion control that best suits a customer environment. i.e. High speed low latency etc. |
Fast Start |
Connection throughput increases very early into the connection |
High Speed TCP congestion control |
Delayed ACK |
Speed up packet transmission by acknowledging multiple packets at a time |
Reduce roundtrips by Ack’ing multiple packets at a time |
Data De-Duplication |
Reduces the amount of data sent across the network virtually expanding the WAN. Serves data to users at LAN speeds from disk cache. |
Repetitive data patterns are recognized and stored in a cache. When these patterns are seen on the network they are replaced with a “symbol” and reconstructed on the other side of the WAN. |
Compression
LZO |
Reduces the amount of data sent across the network virtually expanding the WAN |
All non repetitive data patterns are compressed, this interacts with WAN memory and creates even greater reduction and speed benefits. |
Bi-Directional Data Store Cache |
Data stored in the cache is immediately available for matching in both directions. Opening and saving files is faster |
Data is stored one and is available for matching and reducing repetitive data in both directions. |
Persistent Data Store Cache |
Cache data is available after power outage or reboot |
Data is stored on disk along with the index. |
Universal Data Store Cache |
No multiple copies of the same data, better scaling of disk storage |
Data is stored once in the cache and available to all peers. No cache segmentation is needed |