Storage switch mirrored write sequence count management

1. A method for writing to a mirrored target, comprising:

receiving a data unit at an ingress processing unit and multi-casting a message to a first egress processing unit and a second egress processing unit, said message including said data unit and sequencing information;

providing said data unit with a first sequence number to a first data store comprising said first egress processing unit calculating said first sequence number based on sequencing information and a transfer size for said first data store, and said first egress processing unit sending said first sequence number and said data unit to said first data store; and

providing said data unit with a second sequence number to a second data store comprising said second egress processing unit calculating a second sequence number based on sequencing information and a transfer size for said second data store, and said second egress processing unit sending said second sequence number and said data unit to said second data store, wherein

said receiving a data unit further includes said ingress processing unit determining a value of a running frame count at a beginning of a sequence from an initiator;

said sequencing information includes said value of said running frame count at said beginning of said sequence from said initiator and a data frame sequence count for said data unit;

said first egress processing unit calculating said first sequence number includes said first egress processing unit calculating a value of said running frame count at a beginning of a sequence for said first data store;

said first sequence number is calculated by said first egress processing unit by subtracting said value of said running frame count at said beginning of said sequence for said first data store from said value of said running frame count at said beginning of said sequence from said initiator and adding said data frame sequence count for said data unit;

said second egress processing unit calculating said second sequence number includes said second egress processing unit calculating a value of said running count at a beginning of a sequence for said second data store; and

said second sequence number is calculated by said second egress processing unit by subtracting said value of said running count at said beginning of said sequence for said second data store from said value of said running frame count at said beginning of said sequence from said initiator and adding said data frame sequence count for said data unit.

2. A method according to claim 1 , wherein:

said data unit is in a Fiber Channel IU.

3. A method according to claim 1 wherein:

said data unit is in an iSCSI PDU.

4. A method according to claim 1 , further comprising:

sending a write command to said first data store and said second data store;

receiving a first reply from said first data store and a second reply from said second data store, said first reply indicates a first transfer size, said second reply indicates a second transfer size; and

informing said initiator of said first transfer size or said second transfer size, whichever is smaller.

5. A method according to claim 1 , wherein:

said calculating said first sequence number comprises subtracting a value of a running frame count at a beginning of a sequence for said first data store from a value of said running frame count at a beginning of a sequence from an initiator and adding a data frame sequence count for said data unit; and

said calculating said second sequence number comprises subtracting a value of a running frame count at a beginning of a sequence for said second data store from said value of said running frame count at said beginning of said sequence from said initiator and adding said data frame sequence count for said data unit.

6. A method according to claim 1 , wherein:

said steps of receiving a data unit, providing said data unit with a first sequence number and providing said data unit with a second sequence number are performed without buffering said data unit.

7. A method according to claim 1 , wherein:

said steps of receiving a data unit, providing said data unit with a first sequence number and providing said data unit with a second sequence number process said data unit at wire speed.

8. A method for writing to a mirrored target, comprising:

receiving a data unit for storage on a first data store, said first data store mirrors a second data store;

receiving sequencing information for said data unit;

determining a sequence number for said data unit relative to said first data store, said sequence number is based on said sequencing information; and

providing said sequence number and said data unit to said first data store, wherein said determining a sequence number includes subtracting a value of a running frame count at a beginning of a sequence for said first data store from a value of said running frame count at a beginning of a sequence from an initiator and adding a data frame sequence count for said data unit.

9. A method according to claim 8 , wherein:

said sequencing information includes said running frame count, said value of said running frame count at said beginning of said sequence from said initiator and said data frame sequence count for said data unit.

10. A method according to claim 9 , wherein:

said steps of receiving a data unit, determining a sequence number and providing said sequence number are performed by an egress processing unit associated with said first data store;

said data unit is received from an ingress processing unit;

said data frame sequence count for said data unit was received at said ingress processing unit from said initiator and sent to said egress processing unit with said data unit; and

said value of said running frame count at said beginning of said sequence from said initiator is calculated by said ingress processing unit.

11. A method according to claim 8 , wherein:

said steps of receiving a data unit, determining a sequence number and providing said sequence number are performed without buffering said data unit.

12. A method according to claim 8 , wherein:

said steps of receiving a data unit, determining a sequence number and providing said sequence number process said data unit at wire speed.

13. A method according to claim 8 , further comprising:

sending a write command to said first data store and said second data store;

receiving a first reply from said first data store and a second reply from said second data store, said first reply indicates a first transfer size, said second reply indicates a second transfer size; and

informing said initiator of a burst size, said burst size is said first transfer size or said second transfer size, whichever is smaller;

wherein said data frame sequence count for said data unit is based on said burst size.

14. One or more processor readable storage devices having processor readable code embodied on said one or more processor readable storage devices, said processor readable code for programming one or more processors to perform a method comprising:

receiving a data unit for storage on a first data store, said first data store mirrors a second data store;

receiving sequencing information for said data unit;

determining a sequence number for said data unit relative to said first data store, said sequence number is based on said sequencing information; and

providing said sequence number and said data unit to said first data store, wherein said determining a sequence number includes subtracting a value of a running frame count at a beginning of a sequence for said first data store from a value of said running frame count at a beginning of a sequence from an initiator and adding a data frame sequence count for said data unit.

15. One or more processor readable storage devices according to claim 14 , wherein:

said steps of receiving a data unit, determining a sequence number and providing said sequence number are performed without buffering said data unit; and

said steps of receiving a data unit, determining a sequence number and providing said sequence number process said data unit at wire speed.

16. An apparatus for writing to a mirrored target, comprising:

one or more ports;

one or more processing units in communication with said one or more ports, said one or more processing units receive data units that are to be stored in multiple data stores and translate sequence counts in said data units differently for different data stores accepting different transfer sizes, said one or more processing units translate sequence counts by updating an initial sequence count with one or more new sequence counts; and

said one or more new sequence counts for a particular data unit are calculated by subtracting a value of a running frame count at a beginning of a sequence for a particular data store from a value of said running frame count at a beginning of a sequence from an initiator and adding a data frame sequence count for said particular data unit.

17. An apparatus according to claim 16 , wherein:

said one or more processing units include an ingress processing unit, a first egress processing unit and a second egress processing unit;

said ingress processing unit is in communication with an initiator, said first ingress processing unit receives a first data unit from said initiator and multi-casts a message with said first data unit to said first egress processing unit and said second egress processing unit, said message includes a running frame count, said message includes a value of said running frame count at a beginning of a sequence from said initiator and a sequence count from said initiator;

said first egress processing unit is in communication with a first data store, said first egress processing unit translates said sequence count from said initiator by subtracting a value of said running frame count at a beginning of a sequence for a first data store from said value of said running frame count at said beginning of said sequence from an initiator and adding said sequence count from said initiator; and

said second egress processing unit is in communication with a second data store, said second egress processing unit translates said sequence count from said initiator by subtracting a value of said running frame count at a beginning of a sequence for a second data store from said value of said running frame count at said beginning of said sequence from an initiator and adding said sequence count from said initiator.

18. An apparatus according to claim 17 , wherein:

said first egress processing unit sends a first write command to said first data store;

said second egress processing unit sends said first write command to said second data store;

said first data store replies to said first write command with a first transfer size;

said second data store replies to said first write command with a second transfer size; and

said ingress processing unit receives said first transfer size and said second transfer size and returns whichever is smaller to said initiator so that said initiator can use said smaller of said first transfer size and said second transfer size to create said sequence count from said initiator.

19. An apparatus according to claim 16 , wherein:

said data units can be within packets in accordance with an iSCSI protocol or in accordance with a Fiber Channel protocol.

20. An apparatus according to claim 16 , wherein:

said one or more processing units are capable of processing said data units without buffering.

21. An apparatus according to claim 16 , wherein:

said one or more processing units are capable of processing said data units at wire speed.

22. An apparatus for writing to a mirrored target, comprising:

one or more processing units; and

one or more ports in communication with said one or more processing units, said one or more port are in communication with an initiator and at least two storage devices;

said one or more processing units receive a data unit from said initiator, calculate a first sequence number for said data unit with respect to a first storage device of said at least two storage devices and calculate a second sequence number for said data unit with respect to a second storage device of said at least two storage devices; and

said data unit and said first sequence number are provided to said first storage device via said one or more ports, and said data unit and said second sequence number are provided to said second storage device via said one or more ports, wherein said first sequence number is calculated by subtracting a value of a running frame count at a beginning of a sequence for said first data store from a value of said running frame count at a beginning of a sequence from said initiator and adding a data frame sequence count for said particular data unit; and

said second sequence number is calculated by subtracting a value of a running frame count at a beginning of a sequence for said second data store from said value of said running frame count at a beginning of a sequence from said initiator and adding said data frame sequence count for said particular data unit.

23. An apparatus according to claim 22 , wherein:

said one or more processing units include an ingress processing unit, a first egress processing unit and a second egress processing unit;

said ingress processing unit is in communication with an initiator, said first ingress processing unit receives a first data unit from said initiator and multi-casts a message with said first data unit to said first egress processing unit and said second egress processing unit, said message includes a running frame count, a value of said running frame count at a beginning of a sequence from said initiator and a sequence count from said initiator;

said first egress processing unit calculates said first sequence number by subtracting a value of said running frame count at a beginning of a sequence for said first data store from said value of said running frame count at said beginning of said sequence from said initiator and adding said sequence count from said initiator; and

said second egress processing unit calculates said second sequence number by subtracting a value of said running frame count at a beginning of a sequence for said second data store from said value of said running frame count at said beginning of said sequence from said initiator and adding said sequence count from said initiator.

24. An apparatus according to claim 23 , wherein:

said first egress processing unit sends a first write command to said first data store;

said second egress processing unit sends said first write command to said second data store;

said first data store replies to said first write command with a first transfer size;

said second data store replies to said first write command with a second transfer size; and

said ingress processing unit receives said first transfer size and said second transfer size and returns whichever is smaller to said initiator.

25. An apparatus according to claim 22 , wherein:

said data unit can be within a packet in accordance with an iSCSI protocol or in accordance with a Fiber Channel protocol.

26. An apparatus according to claim 22 , wherein:

said one or more processing units are capable of processing said data unit without buffering.

27. An apparatus according to claim 22 , wherein:

said one or more processing units are capable of processing said data unit at wire speed.