Response to "SIP And H.323 For Voice/Video Over IP -- Complement, Don't Compete!"

From: Paul Long [mailto:plong@packetizer.com]
Sent: Friday, August 03, 2001 12:26 PM
To: elid@tlv.radvision.com
Subject: SIP And H.323 For Voice/Video Over IP -- Complement, Don't Compete!

Eli,

Great article! Almost exactly my sentiments. I have a few bones to pick, however.

1. "Many functions in a SIP-based network rely upon complementary protocols, including IP."
Maybe your original remark got mangled by the editor, but of course H.323 also relies "upon complementary protocols, including IP."

2. "A SIP architecture requires a proxy server to route calls to other entities and a registrar."
I don't believe this is true. Neither H.323 or SIP actually require a gatekeeper/proxy--they both allow endpoints/user agents to contact each other directly. Maybe you meant something else, though.

3. With regard to H.323: "New codes have to have an agreed-upon definition between parties prior to a call."
This is true for all protocols, including SIP. The meaning of codepoints, in whatever form, must be agreed upon by all relevant parties. Again, maybe I am missing your point.

4. With regard to SIP: "Numbers don't have to be allocated to commands for each message in advance."
I don't understand what you are saying here. Although, numbers, per se, do not have to allocated in SIP (well, what about response codes like 200?), codepoints do, just like with H.323, et al. SIP simply uses textual lexemes for codepoints whereas H.323 uses binary values.

5. SIP: "If text commands are added, the other side automatically understands them."
Uh, no. :-) What do you mean by this?

6. "...SIP is more flexible and scalable..."
Just like, "SIP is simple," I know that this is the conventional wisdom these days, but it is not true.

Paul Long