How Does Argent Automatically Optimize TCP/IP?
Argent does a lot behind the scenes to automatically optimize TCP/IP performance for you.
Here are some of the more important ones.
The timing for periodical connections to the TCP/IP server is fully randomized. In other words, if an Argent product needs to connect to an Argent service over TCP/IP once a hour to send data, then the minute of the hour is randomized – could be the first minute or it could be the 59th minute of the hour.
Examples are the replication of the shared database for Argent Console Engine and downloading the control information to the Argent Guardian regional scheduling engine.
This technique prevents the remote clients or regional scheduling engines from trying to connect to the central engine simultaneously and saturate the channel.
A 64-bit CRC is used for every request packet and reply packet. This prevents corrupted data from being transmitted. Though the design of TCP guarantees the integrity of the packet, the implementation of TCP is from Microsoft…
(We still find once in a while, random corruption shows up when transmitting a big chunk of data such as master catalog, a shared database or the Argent Guardian control information.)
By adding this additional layer of bullet-proofing we prevent these horrific errors from affecting you; it’s typically one in 10**9 packets can be out of sequence.
We also change the seed for encrypting request and reply packets. As the encryption seed is different for each packet, deciphering the packets are much more difficult.