Wireshark · Wireshark-dev: Re: [Wireshark-dev] Using Google Protobuf to Export Full Packet Dissection Data via Named Pipe

[Pascal Quantin <pascal.quantin@xxxxxxxxx>]

Hi Mark,

2017-07-11 16:07 GMT+02:00 Mark@xxxxxxxxxxx <mlandri@xxxxxxxxxxx>:

Thanks Roland! 

I guess I'm asking if it'd be value added for me to submit my protobuf solution as an addition to current Wireshark dev branch. I've already written the code. I'd just have to figure out how to incorporate it into the Wireshark build process. It's written in c++ and requires pthread and protobuf libs be installed.

This is a first issue as all Wireshark code (except the Qt GUI) is written in C and this is not something we discussed changing so far.

Happy to do it but would be good to know beforehand if it'd be compatible with Wireshark design ethos and if the community would see value in it.

Sent from my iPhone

On Jul 11, 2017, at 9:00 AM, Roland Knall <rknall@xxxxxxxxx> wrote:

Did you take a look at tshark's -T parameter? "tshark -T jsonraw" for instance, delivers full dissection in Json format. What would be needed is only to shove that into a pipe to capture from some other place.

Cheers

Roland

On Tue, Jul 11, 2017 at 2:48 PM, Mark Landriscina <mlandri@xxxxxxxxxxx> wrote:

Apologies in advance if this question is a bit long-ish.

I've been wondering why Wireshark/tshark doesn't offer the option to export full packet dissection data via named pipe (serialized binary data). Is this due to design philosophy, lack of offers to write the code, or some other reason? Of course, packet dissection data can be written out to stdout or a file in xml format. Perhaps this meets most needs?

Reason for the question is that I needed a dissection data export option that was more efficient than xml. My solution was to modify tshark so it can leverage Google Protocol Buffers to export packet dissection data as serialized binary data. Serialized dissection data is written out to a named pipe. Protobuf dissect tree creation, serialization, export code is all written in C++ and takes advantage of all the optimization work Google has put into its Protobuf library. The client/read side of the pipe can be written in any language supported by the Protobuf library. I wrote mine in Python. The client reads and parses the serialized dissection data (again) using Google Protobuf lib recreating dissection tree data on client side.

Would it be advantageous to incorporate the above Protobuf approach into the Wireshark project or would the community consider it unnecessary or perhaps undesirable?

If you're curious about implementation, you can see my project at the following location: https://gitlab.com/MLandriscina/protoShark.git. This is the first time that I've used Protobuf, so I wouldn't be surprised to discover that better implementations are possible.

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request@wireshark.org?subject=unsubscribe

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
            mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request@wireshark.org?subject=unsubscribe