FAQ: What is the function of the initparm parameters tcp.crc and tcp.crc.override in the initparms?
1.1 Strenghen Data Integrity over TCP/IP Networks (CRC Checking)
Because of customer requests, CRC checking was added to Sterling Connect:Direct for Windows. A CRC will be calculated for each RU / Packet to determine if any change to the data has occurred while the data was in transit.
1.1.2 Operational Characteristics
To accomplish the Cyclic Redundancy Check (CRC), IBM Sterling Connect:Direct will create an optional 32-bit CRC for each RU / Packet sent by the Pnode. This CRC value will be calculated on the entire buffer, including the header. The CRC will be placed in our TCP header prior to issuing the send to the Snode.
If the Snode is capable of doing CRC, the Pnode will control whether CRC is performed. The Snode will do the calculation of a CRC after successfully receiving a data buffer. Then a CRC check will be performed by comparing the new value to the value received in the TCP header. If the values do not match, an exception response with a sense code will be sent to the Pnode. Then the appropriate message will be issued to the log. This will be an indication that the data has changed in some form since the sending node sent the data.
Once the Pnode receives an exception response with that sense code, the process will be terminated. The appropriate message will be issued and the process placed in timer retry (WS TIMER) status for execution later. An SNMP trap will be generated and sent to a Network Manager (NM) to alert the NM of this serious situation.
The option to do CRC will not be performed automatically but can be globally activated or deactivated using a new C:D initialization parameter, CRC = ON or OFF. An option in the initialization parameter will restrict or allow process execution overrides of the initial setting for CRC.
If CRC overrides are allowed via the initialization parameter, the CRC option can be activated or deactivated using a new process keyword, CRC. If the user codes the CRC keyword in the process but their authorization level does not allow them to perform this task, the process will not be submitted.
Again, if overrides are allowed, the initial setting for CRC can also be overridden on a node-by-node basis via a new Remote Node netmap keyword, CRC.
BecauseThe CRC will be performed only for TCP/IP processes and will not be used for any process using Secure+ or SNA which already provides a level of security. If CRC is activated for such an SNA or Secure+ processes, the CRC will be ignored.
BecauseThe Snode will be notified that CRC is being performed via a new FMH68 key, PCRC. If the Snode is capable of doing CRC, the FMH70 key, PCRC, should be turned on. When the Pnode receives an FMH70 back with the notification turned off, then CRC will not perform CRC for that process.
The Copy Termination Stat record (CCTR) will contain an indication that C:D attempted CRC and if CRC was performed for the process ran.
Diagnostic traces will display the generated CRC on both the Pnode and Snode.
Sterling Connect:Direct for Windows Server
FAQ: What is the function of the initparm parameters tcp.crc and tcp.crc.override in