Moved to sip-implementors list ...

Most sip stacks that have wide interoperability have many person years of
engineering effort in them. Even if you had a perfect stack today, it is a
significant ongoing effort to keep it tracking the changes and keep it up to
date. It is hard to answer questions like the one you ask because the
answers will vary widely. I will take a stab at this but keep in mind these
are very rough and would probably vary by two orders of magnitude based on
what you wanted to do.

SIP looks surprisingly easy to implement - and you can get something that
makes calls working very quickly. I'm will to guess that Henning has had
graduate students that wrote working SIP applications in a time span of a
few days to a week. On the other end  several large company have many
engineers long term dedicated to having working SIP stacks. Many vendors
have found it cheaper to license telecom stacks from others than develop
there own - check out the number of people using other vendors H.323 or SS7
stacks.

If you want to get a feel of whats in a SIP stack, check out some of the
open source ones such as the one at NIST or www.vovida.org.

To build a SIP stack you need competent programmers that can do network
programming, protocol parsing and such. It also helps greatly to have a few
people who deeply understand how SIP works - why it works that way - and
what the complex things are that have messed up other implementation.
Ironically as SIP has gotten more complex and more widespread recognition,
the market has changed in a way that's makes it easier to hire these sort of
people than ever before.

A team of 5 engineers could likely put together an initial working
implementation suitable for a complex IP phone in 3-5 months and after a
sipit or two have it widely interoperable. A proxy implementation is
simpler. After the initial 4-9 months a team of 2 engineers could likely
support it and keep it tracking the standards. Keep in mind the accuracy of
these estimates is likely in the multiple or divide by 50 zone. What
features of SIP you need, how interoperable the implementation is, what you
do with security, how much previous experience you have, how small or fast
you needed it, and so on all change this estimate a bunch.

As a final question I often ask people who are about to implement their own
SIP stack - what TCP stack are you going to use? I have never had some tell
me there were writing their own. In fact, they often say they will us one
that can track it's roots back to a BSD stack. It always makes me wonder -
why not use an open source stack. There are a few around. It may be that one
of the open source ones is lacking some feature you needs but, would it not
be easier to add that feature than write a complete new stack? What's your
competitive advantage - hopefully it's your application - people who think
their SIP stack is their sustainable competitive advantage you are probably
unclear on the goals of SIP.

Cullen

> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of sumir
> Sent: Monday, November 25, 2002 11:11 PM
> To: [EMAIL PROTECTED]
> Subject: [Sip] Sip Stack Market Place. Follow-Up!
>
>
> Hi friends
>
> I am highly thankful to Mr. Mohammed Asfar of Nokia and Mr. Jeff
> Hoefgen of
> Radvision for their answers...
>
> I hv some questions like:
> Q1 What if I start developing a sip stack. what techical knowledge and
> resources I require?
> Q2 What is the approximate development time for developing sip stack? (man
> days).
>         Q3
> How many people is needed to dedicate to developing sip stack?
> Q4 What are my costs for developing these stacks?
> Q5 What technical resources I need to build these stacks?
> Q6 Do i need to follow certain rules if any..
>
> Looking forward to hear from you friends soon
>
> With regards!
> sam
>
> >

_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to