Little DICOM inconformance - Maximum Length Received subitem
Forums > Onis operations > Others
|
|
|
|
| mulOMpUR | Hi guys,
This is definitely non-critical.
Onis version: 2.3.5
Our side: COALA system (under implementation)
We support any value for Maximum Length Received. We are testing our server with varous configurations.
We get some problems when trying to get image on demand using Onis 2.3.5 and specifying 0 as Maxumum Length Received upon association establishment.
Zero Maximum Length Received is correct value in the standard and means that the maximum length is not specified (so other DICOM AE can use any value it finds feasible)
Excerpt from DICOM PS 3.8-2011 Annex D:
"This parameter allows the association-requestor to restrict the
maximum length of the variable field of the P-DATA-TF PDUs sent
by the acceptor on the association once established. This length
value is indicated as a number of bytes encoded as an unsigned
binary number. The value of (0) indicates that no maximum length
is specified. This maximum length value shall never be exceeded
by the PDU length values used in the PDU-length field of the
P-DATA-TF PDUs received by the association-requestor.
Otherwise, it shall be a protocol error."
Upon establishing association Onis answers with A-ASSOCIATE-AC without Maximum-Length-Subitem.
This is mandatory item in DICOM v3 and any DICOM v3 application can expect to receive this subitem.
(it is safer not to assume that howerver for bacwards compatibility)
Excerpt from DICOM PS 3.8-2011 Annex D:
"The Maximum Length Item support is required for all DICOM V3.0 conforming implementations."
So to conclude:
Onis should insert it's own maximum length received in A-ASSOCIATE-AC, either 0 or some other value.
It does that when we specify non-zero maximum length received.
One thing that seems suspicious is the fact that Onis always (apart from 0 case) mimics the value it receives in Maximum-Length-Subitem. That may be naturally intentional.
Kind Regards,
Bartosz Meglicki
National Centre for Nuclear Research
Department of Nuclear Equipment
ul. Andrzeja Sołtana 7
05-400 Otwock, Poland
| | 11/8/2011 11:06:27 PM |
| mulOMpUR | I have thought about that a little more.
Dicom PS 3.7-2011 D.3.3.1 (Maximum Length PDU Notification) also states that "This notification is required".
So it indeed seems to be mandatory.
I have some difficulties tracking whether it was mandatory in previous versions of the standard (before v3)
Perhaps the safest way to implement it is to always reply with Maximum Length Subitem if it was present in A-ASSOCIATE-RQ and not including it if it was not present. Also ability to accept A-ASSOCIATE-RQ/AC without that field seems reasonable to allow communication with non-conformant servers.
B.M.
| | 11/9/2011 6:46:56 AM |
| mulOMpUR | Hello Bartosz,
If you want to try the new version, this was fixed as well. Let me know!
Regards,
madric
| | 11/9/2011 5:43:07 PM |
| mulOMpUR | Hi Madric,
Certainly, as I mentioned in some other post we are already testing our radiotherapy COALA system with Onis and other DICOM applications.
Naturally, we are always interested in testing with the most recent development version.
Regards,
Bartosz Meglicki
National Centre for Nuclear Research
Department of Nuclear Equipment
ul. Andrzeja Sołtana 7
05-400 Otwock, Poland
| | 11/9/2011 8:21:44 PM |
| mulOMpUR | Hi,
I've just sent an email to you with a link for downloading a new version. I think it will be fine for you now.
Also for information, about this:
"One thing that seems suspicious is the fact that Onis always (apart from 0 case) mimics the value it receives in Maximum-Length-Subitem. That may be naturally intentional. "
This is intentional. As written in the conformance statement of Onis:
2.1.1.1 General
The AE will accept any PDU length, offered on associations initiated by remote applications. However, the actual PDU size will not exceed the maximum size requested by the SCP. It is recommended that it does not exceed 64 Kbytes.
Warm regards,
madric
| | 11/9/2011 8:33:31 PM |
| mulOMpUR | Hi Madric,
Oh, ok. I admit I haven't looked into Onis conformance statement.
Just curious, why reply with the same value?
In DICOM max pdu length received is something each application reports upon association establishment. The meaning is - you can send anything that is lower.
So application A says in ASSOCIATE-RQ - we can handle up to X bytes.
Applicaion B replies in ASSOCIATE-AC - we can handle up to Y bytes.
Is it for perfomance reasons? E.g. "if you can receive up to X bytes send also the same amount to us if you like"?
Kind regards,
B.M.
| | 11/9/2011 10:16:43 PM |
| mulOMpUR | Hello Bartosz,
at the time we wrote the association code, we just had a look to the conformance statement of other DICOM products and find out that some of them were using this rule. We simply decided to use the same one.
Kind regards,
madric
| | 11/10/2011 10:17:05 AM |
|
|
|
|
| |
|