RMAN Archive Log Backup

We are having an intermittent problem related to failing RMAN archive log backups. Normally everything works fine but once every fortnight or so the archive log backup starts to fail with the below message. We run two Oracle instances on this windows box, so they are using same oracle and TSM client binaries, but it is only of the instances that have this problem - the largest instance though. Restarting the Oracle instance solves the problem and everything is back to normal, until next time ...

This is the error message in the rman log file
CODERMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03009: failure of allocate command on c1 channel at 03/27/2013 11:00:29
ORA-03113: end-of-file on communication channel

Alert log contains this
CODEWed Mar 27 11:00:27 2013
Exception [type: ACCESS_VIOLATION, UNABLE_TO_READ] [ADDR:0xFFFFFFFFFFFFFFFF] [PC:0x7716324E, 000000007716324E]
Errors in file R:\ORA_DATA\SDC002\diag\rdbms\sdc002\sdc002\trace\sdc002_ora_12328.trc  (incident=523573):
ORA-07445: exception encountered: core dump [PC:0x7716324E] [ACCESS_VIOLATION] [ADDR:0xFFFFFFFFFFFFFFFF] [PC:0x7716324E] [UNABLE_TO_READ] []
Incident details in: R:\ORA_DATA\SDC002\diag\rdbms\sdc002\sdc002\incident\incdir_523573\sdc002_ora_12328_i523573.trc

Trace file
CODE
Trace file R:\ORA_DATA\SDC002\diag\rdbms\sdc002\sdc002\trace\sdc002_ora_12328.trc
Oracle Database 11g Release 11.2.0.3.0 - 64bit Production
Windows NT Version V6.1 Service Pack 1
CPU : 40 - type 8664, 20 Physical Cores
Process Affinity : 0x0x0000000000000000
Memory (Avail/Total): Ph:85656M/147445M, Ph+PgF:96175M/157683M
Instance name: sdc002
Redo thread mounted by this instance: 1
Oracle process number: 57
Windows thread id: 12328, image: ORACLE.EXE (SHAD)

*** 2013-03-27 11:00:27.255
*** SESSION ID:(530.31175) 2013-03-27 11:00:27.255
*** CLIENT ID:() 2013-03-27 11:00:27.255
*** SERVICE NAME:(SYS$USERS) 2013-03-27 11:00:27.255
*** MODULE NAME:(rman.exe) 2013-03-27 11:00:27.255
*** ACTION NAME:(0000001 STARTED1) 2013-03-27 11:00:27.255

Exception [type: ACCESS_VIOLATION, UNABLE_TO_READ] [ADDR:0xFFFFFFFFFFFFFFFF] [PC:0x7716324E, 000000007716324E]

*** 2013-03-27 11:00:27.473
Incident 523573 created, dump file: R:\ORA_DATA\SDC002\diag\rdbms\sdc002\sdc002\incident\incdir_523573\sdc002_ora_12328_i523573.trc
ORA-07445: exception encountered: core dump [PC:0x7716324E] [ACCESS_VIOLATION] [ADDR:0xFFFFFFFFFFFFFFFF] [PC:0x7716324E] [UNABLE_TO_READ] []

I have the incident file as well if that is of any use.

Could the problem be related to running two backups in parallel?

We run full backups once a week and both this time and last time the archive log backup failed it was in conjunction to when the full backup also was running.

Is there any other solution to the problem than restarting the oracle instance?

Since this is a busy production instance it is a big problem for us to restart it outside of maintenance windows.