DNS effort summary (Was: Update on Dns server/client ?)

[Stefano Bagnara <ap...@bago.org>]

James Im ha scritto:
> Do you know what is the status of the dns server/client based on mina?

I planned to write an (always delayed) update in the past days, so let's
"pong" you.

Here is my understanding about people opinions:
-----------------------------------------------

1) MINA PMC decided not to "incubate" a similar effort, pointing us to
ASF Incubator or to an ASF Lab (no official statement has been done by
them after they discussed this internally, but I think I understood this
from the last comments).

2) Brian Wellington (dnsjava author) has no problem if dnsjava code is
taken under the BSD and a new community-managed library is produced, but
he won't probably go as far as to submit a software grant to the ASF (he
didn't reply to my specific question). I'm not sure at all we can start
an Incubator project based on code we get from BSD and for which we
don't have the copyrights (and cannot give a software grant).

3) Rob Butler is willing to be an "expert" for a similar effort and he's
already working on a Java5 (not MINA and not SEDA if I understood it
right) implementation. He plan to publish his code under BSD (could be
ASLv2, too).

4) Richard Wallace likes the idea of using Java5+MINA2 and start from
refactoring dnsjava. He suggested we can have a core that is mina
agnostic and a bridge to mina as a reply to Alex D (even if there is an
issue with "MINA ByteBuffer" dependency).

5) James Im likes a Java5+MINA2 approach. Doesn't care of details,
probably more inclined in using the directory-dns-library as a start.

6) Alex D. (dnsjnio author) (dnsjnio is an add-on to dnsjava adding SEDA
/NIO based support for asynchronous resolutions) wants a dnsjava based
library with no third party dependencies (no MINA). His company is
willing to switch the current licensing to something more reusable like
ASLv2/BSD, but the code is currently MPL.

7) Michael Mealling is working on a custom dns server based on code from
 dnsjava and dnsjnio: he told his client he's willing to publish the
code to open source sooner or later. I've not understood if his code is
MINA based or not and what level of involvement he's interested in.

8) Julien Vermillard started working on directory-dns project to remove
directory dependencies and make it functional for a basic lookup. He
didn't tell much more to understand his goals and desired involvement.

9) Norman Maurer suggested to start from dnsjava+dnsjnio and get an
Apache Lab to do the work.

10) Stefano Bagnara (me ;-)) likes the idea of starting from
dnsjava+dnsjnio, merge them and apply some step-by-step refactoring. The
goal would be to keep the core mina-free but to have a good bridge to
mina (for both the executors/monitoring features in mina and to have a
standard DNS client for mina).

Feel free to reply to this to correct, complete informations or to add
your opinion.

Now some personal consideration:
--------------------------------

A. I don't think an Apache Lab would work because it requires everyone
to be already an ASF committer (and this is not the case of most
mentioned people) and it is specific to new code, and as I said I'd
start with dnsjava+dnsjnio.

B. About ASF incubator I guess the only one that could do something
would be Brian. Correct me if this is wrong and we can incubate dnsjava
without a software grant from him.

C. As MINA PMC didn't offer to sandbox such an effort (and I don't even
know if they're allowed under the ASF rules) I think ASF has no more
options for such an effort.

And my personal conclusion:
---------------------------

I'm not sure I see a solution to this stall: maybe we should create a
google code project and a google group (or any other code hosting
provider) with every of the above people subscribed and with write
access to the repository.

I'm not even sure I can really identify a common goal in the 9 opinions
recorded, but not having a repository to write code in won't ever give
us a solution. If there is consensus between the above people I can
create the google infrastructure and we can continue our discussions and
more important start coding something there.

Stefano

Re: DNS effort summary (Was: Update on Dns server/client ?)

[Michael Mealling <mi...@refactored-networks.com>]

My server is MINA 1.1.0 based. What it does is receive an ENUM request 
(RFC 3761) and chops off the root domain. Then it looks in its config 
for a list of VOIP peer ENUM roots and creates a new query for the 
original TN with each peer. It aggregates all of the results, renumbers 
the priority values according to the configuration, and sends all of 
that back to the client.

So its fairly specific to ENUM. But I does have both TCP and UDP 
listeners, SNMP4J integrated so it behaves nicely in an enterprise 
nework management environment, has a reasonable performance statistics 
collection system, and uses dnsjnio to handle the asynchronous queries 
for all of the VOIP peers. I personally like using dnsjava inside the 
protocol encode/decode bits since for me it looks as though the entire 
protocol level functionality is handled by a handful of lines in the 
codec classes. It just felt elegant.

There _may_ be a chance that I could donate the entire thing to ASF. 
Moer on that later...

-MM

Stefano Bagnara wrote:
> 7) Michael Mealling is working on a custom dns server based on code from
>  dnsjava and dnsjnio: he told his client he's willing to publish the
> code to open source sooner or later. I've not understood if his code is
> MINA based or not and what level of involvement he's interested in.