Method of delaying checkpoints by inspecting network packets

1. A method of delaying an initiation of a checkpoint in a virtual machine system comprising the steps of:

establishing an input/output module;

buffering a deferrable or non-deferrable frame in the input/output module;

determining whether the frame is a deferrable frame or a non-deferrable frame;

if the frame is a deferrable frame, delaying an initiation of a checkpoint; and

not transmitting the frame until the checkpoint occurs.

2. The method of claim 1 wherein the frame is examined to determine if it is a deferrable frame.

3. The method of claim 2 wherein the deferrable frame is a packet tagged with one or more of:

not being a last data fragment of a multi-fragment message;

a data segment that does not indicate a flush of a send buffer; and

a segment that only serves as an acknowledgment of some prior message.

4. The method of claim 2 wherein the deferrable frame is an internet protocol (IP) packet tagged with one or more of:

a ‘more fragments’ attribute;

a transmission control protocol (TCP) data segment that lacks a push (PSH) flag and that carries no flags other than an acknowledgement (ACK); and

a transmission control protocol (TCP) segment that contains no data and carries only the acknowledgement (ACK) flag.

5. The method of claim 2 wherein the deferrable frame is identified by a protocol in use.

6. The method of claim 5 wherein the deferrable frame is identified by a session-state of the protocol in use.

7. The method of claim 2 wherein the deferrable frame is identified by port number.

8. The method of claim 7 wherein the port number is a source port number.

9. The method of claim 7 wherein the port number is a destination port number.

10. The method of claim 2 wherein the deferrable frame is identified by network address.

11. The method of claim 10 wherein the network address is a source network address.

12. The method of claim 10 wherein the network address is a destination network address.

13. The method of claim 1 further comprising the step of concatenating the checkpoint delays due to a plurality of deferrable frames.

14. The method of claim 13 further comprising the step of setting an upper limit to the amount of delay that can be generated.

15. The method of claim 1 wherein the frame is determined to be deferrable based on a network interface in use.

16. A fault tolerant computing system comprising:

a first computing device comprising a primary virtual machine; and

a second computing device comprising a secondary virtual machine,

wherein the first computing device and the second computing device are networked,

wherein the primary virtual machine comprises a first checkpointing engine and a first network interface having a first input/output module for buffering frames,

wherein the secondary virtual machine comprises a second checkpointing engine and a second network interface,

wherein the primary virtual machine buffers a frame in the first input/output module,

wherein the primary virtual machine determines whether the frame is a deferrable frame or a non-deferrable frame,

wherein the first checkpointing engine delays an initiation of a checkpoint for each of the one or more deferrable frames of data in the first input/output module, and

wherein the deferrable and non-deferrable frames in the first input/output module are not transmitted until the checkpoint occurs.

17. The fault tolerant computing system of claim 16 wherein the first checkpoint engine pauses the primary virtual machine when a checkpoint is declared.

18. The fault tolerant computing system of claim 17 wherein the first checkpoint engine adds an additional delay before declaring a subsequent checkpoint.

19. The fault tolerant computing system 16 wherein the first checkpoint engine captures the state of the primary virtual machine while the primary virtual machine is paused.