Notwithstanding T1 is full-duplex, a frame-relay access will never perform like a dedicated T1, Reason being, you are sharing circuits somewhere with someone else. Try asking your ISP what the CIR on the circuit is and that could give you a better idea of what to expect.
Hope this helps, please rate post if it does!
"We have frame-relay for Internet access,at full T1 speed"
Based on the original post it appears the CIR is a full T1. If the provider guaranteed a full T1 then irrespective of the fact it's a shared circuit they should be getting full T1 bandwidth in both directions.
CIR same as access speed negates the reason itself for using frame-relay: economics, and is unlikely to happen.
But, we can get better detail from the original poster.
"show frame-relay map" is not supposed to relate to bandwidth. But if you do "show frame-relay qos-autosense" and the SP is using Cisco, and certain configurations are made, you will see CIR and other service parameters. Too many "and" perhaps.
While I agree with you that we need more details from the original poster I differ about your view having 1.5 mbps CIR negates the need for frame relay. What if the headend or ISP device has a DS3 frame relay circuit and customer wants to have full T1 bandwidth then frame relay transport is inevitable. It's not always the economics that determines the technology chosen.
LMI type Cisco would discover the CIR set on the provider switch when show frame-relay map command is issued. If you see the CIR as 1.5 mbps then yes that indeed is the CIR of the PVC.
What happens, it not that the customer 'wants' to have frame-relay. He may be presented with two estimates, for example, $1,000 monthly for a T1 to the nearest FR SP POP, or $1,500 for a dedicated T1 to the nearest ISP POP (that is more miles). He is told that performances are comparable. Which one would you choose ?
Most customers don't even -know- what frame-relay is.
One case where FR choice becomes more convenient, again compared to dedicated circuits, is when beside Internet access, there are al PVC connected to branches. But with the advent of VPNs, this is also becoming less and less common.
While I don't come from the business side to know the exact costs associated with each technology I can tell you that economics isn't always the determining factor for so many customers. I have seen so many flourishing businesses use a legacy/outdated technology not because they can't afford to choose a robust alternative but due to some other limitations. Like it or not SNA is still around a good bit.
Going back to the original poster's concern the fact that show frame-relay map shows the bw as 1.5 mbps clearly indicates the dlci is provisioned for a full t1 bandwidth. As you may know Cisco LMI has a field that allows it to learn the bandwidth value set on the switch.
One more comment on this interesting discussion.
CIR equal access-speed doesn't necessarily means guaranteed bandwidth. On the other end of the PVC (DS3 likely), the ISP can be using "CIR overbooking", that is, sums of CIR for all PVCs exceeds access speed. Albeit a questionable practice, it's technically perfectly feasible on all the FR gear that I know.
Note, is not that I have anything against FR per-se, I'm just trying to make clear that it is a shared media and anyone using it should be informed of that.