Imported from previous forum
FIXT and the earlier transports are ordered.
Messages are delivered in the same order that they are sent.
The FIX application messages (that I’ve looked at so far) that flow over the transport look to be defined in such a way that they can be operated over an unordered transport.
Message linkages seem to be used to enable the establishment of order.
For example, a new IOI has a new ID and identifies which ID it supercedes.
Today I expect that any application that receives a message out of order will probably raise an error and fail to process the message.
However, in principle, it seems an application could be written that reconstructs the order.
This would allow the operation of FIX over unordered transports, such as message busses that multithread their routers, or the operation of FIX applications in application servers that multithread their MDB listeners.
I’m not expecting FIXT to ever become unordered, but I do observe that other financial messaging standards referencing unordered transports and ISO 20022-1 part 6.
Q1: Does FIX believe all of its application messages can be operated over an unordered transport?
Maybe I missed a few that don’t work.
Q2: If so, does FIX plan to ensure this continues to be the case, as new messages are defined?
Seems cheap to do, and leaves the door open for the future.
Q3: Does FIX ever plan to define any unordered transports, or declare support for them?
I know people ship FIXML around over products like MQ and EMS, sometimes with brokering involved, and as the infrastructure to do this becomes more sophisticated, I can imagine interest in unordered transports might increase.
{{{ Andy
Please take a look at “High Performance Interface WG” at
http://fixprotocol.org/working_groups/highfrequency
where many ideas for futuristic trading protocol(s) is being discussed. The discussion forum is
http://fixprotocol.org/discuss/149
Hope you find something interesting there.
FIXT and the earlier transports are ordered.
Messages are delivered in the same order that they are sent.The FIX application messages (that I’ve looked at so far) that flow over the transport look to be defined in such a way that they can be operated over an unordered transport.
Message linkages seem to be used to enable the establishment of order.
For example, a new IOI has a new ID and identifies which ID it supercedes.Today I expect that any application that receives a message out of order will probably raise an error and fail to process the message.
However, in principle, it seems an application could be written that reconstructs the order.This would allow the operation of FIX over unordered transports, such as message busses that multithread their routers, or the operation of FIX applications in application servers that multithread their MDB listeners.
I’m not expecting FIXT to ever become unordered, but I do observe that other financial messaging standards referencing unordered transports and ISO 20022-1 part 6.
Q1: Does FIX believe all of its application messages can be operated over an unordered transport?
Maybe I missed a few that don’t work.Q2: If so, does FIX plan to ensure this continues to be the case, as new messages are defined?
Seems cheap to do, and leaves the door open for the future.Q3: Does FIX ever plan to define any unordered transports, or declare support for them?
I know people ship FIXML around over products like MQ and EMS, sometimes with brokering involved, and as the infrastructure to do this becomes more sophisticated, I can imagine interest in unordered transports might increase.
{{{ Andy