Dear Christian,

First of all, thank you for sharing these prototypes. This would be very
helpful for me. However, I encountered some problems in building
the source.

Below is what it did:

I downloaded the code in: https://github.com/CPqD/ofsoftswitch13

Based on README.md, I performed "Building" steps below.
-   $ ./boot.sh
    **** contained some errors below ****
fatal: Not a git repository (or any of the parent directories): .git
         debian/automake.mk:2: trailing backslash on last line
         Makefile.am:67:   `debian/automake.mk' included from here
         autoreconf: automake failed with exit status: 1
-   $ ./configure
    **** contained the following errors ****
         configure: creating ./config.status
         config.status: error: cannot find input file: Makefile.in
After executing above scripts, I could not execute '$ make'.

I have no idea as of yet on how to fix these errors. I tried the following commands: aclocal, autoheader, autoconf, automake ... suggested from websites, but the errors
are still the same. Could you give me some advice on how to proceed.

By the way I'm using CentOS v6.2 environment. Would this be a problem?

Thanks,
Ken




On 3:59, Christian Esteve Rothenberg wrote:
Dear OpenFlow fellows,

giving continuity to the OpenFlow 1.X developments by Ericsson and
CPqD (cf. v1.2 Tool-Kit [1]) we are glad to announce the release of a
version 1.3 software switch [2] (based on Stanford reference design
extended by Ericsson for v1.1) and a companion NOX OpenFlow 1.3
controller [3] (based on NOX Zaku and
using oflib 1.3 of the software switch).

Find below a quick user's guide.

Early users, feedback, and contributions are very welcome!


HOW-TO

- For further instructions on how compile and deploy, please go to the
READMEs of the software switch  [5] and NOXv1.3 [6].
- See the learning switch and discovery apps examples on how to use
oflib v1.3 APIs.
- For quick testing you can use the dpctl switch built-in controller
and watch the wireshark OFPv1.3 control traffic.

CAVEATS

- Performance: The software switch implementation prioritizes OpenFlow
1.3 feature compliance and overall correctness.
- Feature completeness: See pending protocol optional features in the
living TODO list [7]  based on Annex B.11 of the OpenFlow v1.3 spec
[8]. Notably, the policy metering actions are not yet enforced in the
data plane.
- Limitations in dpctl: (no OXM mask support, lack of multipart messages,
limited fields for set_field action)
- Missing APIs in NOX to simplify access to oflib1.3

NEXT STEPS

- Complete implementation missing features (B.11.1, B.11.4, B.11.5,
B.11.6). Please, track the TODO list [7].
- Complete wireshark dissector OFP for newly introduced messages (e.g.
flow metering messages) and some 1.2 pending decoding (e.g. stats
bodies)
- Extend dpctl
- Testing: Interop with OpenFlow v1.3 controllers? OF-Test 1.3?
- Put together a v.1.3 Tool-Kit in pre-configured VM (including
Mininet with 1.3 components)
- Make a github wiki page with detailed tutorials for OF1.X.
- Performance optimization: Q1/2013


REFERENCES

[1] 
https://mailman.stanford.edu/pipermail/openflow-discuss/2012-July/003479.html

[2] https://github.com/CPqD/ofsoftswitch13

[3] https://github.com/CPqD/nox13oflib

[4] https://github.com/CPqD/ofdissector/tree/of13

[5] https://github.com/CPqD/ofsoftswitch13/blob/master/README.md

[6] https://github.com/CPqD/nox13oflib/blob/master/README

[7] https://github.com/CPqD/ofsoftswitch13/blob/master/TODO.md

[8] 
https://www.opennetworking.org/images/stories/downloads/specification/openflow-spec-v1.3.0.pdf


--
Christian Esteve Rothenberg, Ph.D.
Converged Networks Business Unit
CPqD - Center for Research and Development in Telecommunications
Tel. (+55 19) 3705 4479 / Cel. (+55 19) 8193-7087
_______________________________________________
openflow-discuss mailing list
[email protected]
https://mailman.stanford.edu/mailman/listinfo/openflow-discuss


_______________________________________________
openflow-discuss mailing list
[email protected]
https://mailman.stanford.edu/mailman/listinfo/openflow-discuss

Reply via email to