Hi Evan,
Frame Relay networks support IP applications, along with apps riding over many other protocols. And IP is what VoIP is all about. So, they would be compatible, if not in actual application for whatever reason, at least in theory. I think that what you are asking is whether the VoIP gateway mfgrs are, by and large, adapting their wares to be compatible with FR, which means are they actually providing the necessary hardware and software interfaces to support FR? Some do, and some don't <with a qualified YET>, and this is one of those weighted attributes worth tracking. Those who don't, currently, will have to adapt sooner or later if they are to survive, IMO. But most already do.
Frame Relay is perceived to be a more reliable, hence more costly, alternative than the public Internet, and is used as a fallback or safety net by others, when all is not well on the Internet. Still other mfgrs will configure their options to be able to revert to the PSTN, or some other virtual net (cloud), when performance degradation thresholds are exceeded.
This area of congestion uncertainty can be responsible for some very peculiar and nebulous cost and ROI predictions, since the fallbacks are most often more expensive than the public Internet, with some of them measured by the minute or packet. The offshoot to all of this could be one of great irony, where sudden over-subscription (or times of heavy backbone usage) could lead to less profitable returns, if the backbone provisions are not guaranteed and scaled properly. <ouch!>
Regarding the other part to your question, there are many ways to support IP over ATM, both directly and indirectly, and this is taking place right now all over the Internet (probably somewhere over the backbone over which this message is being delivered to SI's Web server). Think of it as a passenger (IP) riding on a high-speed train (ATM).
Since these protocols operate at different layers of the protocol stack (ATM at Layers 1 and ~2; TCP/IP at Layers ~3 and ~4), they need not be married to one another, tho.
While ATM can and does support IP, it also offers a new alternative to providing voice services, in the form of compressed voice sent directly, embedded in native ATM payloads. It differs from TCP/IP in many ways (e.g., ATM is deterministic and TCP/IP is probabilistic), hence, it is perceived to be a lot more complicated by the powers that be, or at least by those who are now vested in TCP/IP in one way or another. Which means almost everyone. Then again, I know some folks who are bred on switched services who say the same things about TCP/IP. Go figure.
GDC and Nortel are among the vendors who are well along in the development of Voice and Telephony over ATM (VTOA) delivery mechanisms, as are many of the other principal WAN players (Ascend, LU, Cisco, Alcatel, etc.) with some of them offering products today which are both proprietary and standards based. Most, if not all, of them will eventually support both protocols.
VTOA is beginning to influence some carriers' (particularly European and some domestic value-added network carriers) buying decisions where next generation backbones are concerned. VTOA, like VoIP, is one to be on the lookout for, although you are not going to find many pure plays in this arena. It's a bit too early right now to predict what the uptake will be.
Frank |
