FAST REFERENCE CODE 1.1

, ,
21

[ original email was from Darshan Khedekar - darshan.khedekar.ext@deutsche-boerse.com ]
Hi,

The XML output generated from the sample input data is well-formed if not valid. You can use any XML compatible browser to see the messages.

The intention of the sample data was to help the members to test their applications. The step-by-step debug output should be of more interest to the members as it prints every step in the data decoding process. The intention of XML output was to give a final snapshot of the decoded output.

Our experience with the XML tools isn’t very satisfactory. They are simple too slow in working with large files. It must be mentioned here that Eurex does not plan to provide the data in XML format in production.

The XML decoded output runs into several MB’s within a few seconds of operation.

Eurex Enhanced Broadcast Solution conforms to FAST 1.1 and to Level 1 of “FAST Session control Protocol 1.1” (approved draft).

Coming to the static templates used by Eurex, considering the complexities with working with the UDP based service Eurex did not want to make life more complex for the members. It was a pragmatic decision to keep the templates static to start with. We have tried to overcome the repetition problem by grouping fields together and declaring them as optional collectively. Once the FAST SCP 1.1 is approved and is digested by the community, Eurex may consider introducing dynamic templates.

I am open to any FAST related queries in general or on Eurex data. I would advise you to approach Eurex Customer support for any other queries on the operation of the service.

Regards
Darshan

FAST1.1 requires a lot more of the impl than 1.0 did, so substantial
effort would have to go into a complete implementation.

I second that!

As EUREX uses static templates (compile time), you can get away with a
small subset implementation of FAST. EUREX made an effort to provide
binary data input and expected output in plain text although their XML
doesn’t pass a simple XML validator.

On the other hand, CME is a lot harder to implement fully as templates
are retrieved at run-time and hence you will need to support all cases.
You can dump the data from the CME new feed but it’s up to you to figure
out if the data you are decoding make sense. I would have loved to see
some expected output like EUREX did.

The FAST doc is difficult to handle at first too adding to the overall
complexity of the task. I greatly underestimated the task :slight_smile: Oh well.

Jeremy

22

The XML output generated from the sample input data is well-formed if
not valid. You can use any XML compatible browser to see the messages.

I meant the XML given for messages definition at the end of the EUREX doc. Some tags are incorrects.

The chap working on EUREX had something done quickly, so the pragmatic approach was welcomed :slight_smile:

Thanks for your reply,
Jeremy

23

[ original email was from Darshan Khedekar - darshan.khedekar.ext@deutsche-boerse.com ]
Hi Jeremy,

Yes indeed there are couple of missing front slashes :). Will get the templates corrected in the next version of the doc.

Thanks for pointing out.
Darshan

The XML output generated from the sample input data is well-formed if
not valid. You can use any XML compatible browser to see the messages.

I meant the XML given for messages definition at the end of the EUREX
doc. Some tags are incorrects.

The chap working on EUREX had something done quickly, so the pragmatic
approach was welcomed :slight_smile:

Thanks for your reply, Jeremy