| CSN.1 integration to TITAN? [message #1768776] |
Sun, 23 July 2017 19:20  |
|
Hi!
Does anyone here have experience in integrating with protocols specified in the CSN.1 syntax, as used for 3GPP TS 44.060 (GPRS RRC/MAC) as well as 44.018 (RR system information rest octets)?
I know that TITAN has no native support for this. But given its background in the cellular industry, I'm sure somebody must have done this before.
If there was some external tool that would take the CSN.1 syntax and generate XML + XSD from it, one could use this with the TITAN XML coder. The same would apply for a tool that could transcode CSN.1 syntax to ASN.1 and transcode the data to BER encoding, a bit similar to what Elemér has suggested multiple times for APER/UPER support.
It seems there are CSN.1 to TTCN-3 converters from Testing Technologies (http://www.testingtech.com/products/ttplugins_csn1.php) and Spirent (http://www.spirent.cn/Products/TTworkbench/TTplugins) but I'd assume they are specific to their TTCN-3 toolchain and they're unlikely to work with TITAN.
Most of the (few) CSN.1 compilers / code generaters out there are able to generate Java/C/C++ code for encoding/decoding, but that of course won't help at all with generating the TTCN-3 data types associated with it
I've been thinking a bit about using Wireshark's intenal library to generate XML or JSON and feed that into TITAN. However, while it can generate XML/JSON dumps of the message, it cannot generate related XSD or JSON schemata. Writing them from hand would probably be just as complex as implementing encoder/decoder from hand :( - and that I'm for sure not prepared to do.
Does anyone have ideas about how to approach this? I'm not fundamentally opposed to licensing a proprietary external tool to do the conversion, as long as it runs on Linux and works with TITAN.
|
|
|
| Re: CSN.1 integration to TITAN? [message #1768827 is a reply to message #1768776] |
Mon, 24 July 2017 08:04  |
|
Hi Harald,
somehow we did not encounter this problem before; probably we did not have to work with these protocols described and encoded with CSN.1.
I don't see any obvious solution to this.
Maybe community can help.
Best regards
Elemer
|
|
|
| Re: CSN.1 integration to TITAN? [message #1768853 is a reply to message #1768827] |
Mon, 24 July 2017 11:05 |
|
Elemer Lelik wrote on Mon, 24 July 2017 08:04
somehow we did not encounter this problem before; probably we did not have to work with these protocols described and encoded with CSN.1.
I don't see any obvious solution to this.
Thanks for your response, though I'm surprised to hear not even Ericsson has solved that problem :P
I'll continue to evaluate some options here. I'll continue to evaluate some options here. It would be a shame to fall back to hand-crafted messages. But even that would be hard to impossible, as CSN.1 doesn't know only the bit values 0 and 1, but also L and H, where L + H are basically relative to the default GSM Um padding 0x2b.
So a bit can not only be defined absolute (like a normal bit in any other context), but relative to whether it is identical to the bit of the normal padding at its position, or whether it is diferent from the nomal padding bit at its position. That's certainly something that's not possible to express with a native TTCN-3 datatype.
|
|
|
|
|
Powered by
FUDForum. Page generated in 0.03095 seconds
] 
Search
Help
Register
Login
Home
