Lol yes... You can stream Pandora on EDGE, you don't need 20mbps down to stream music.
- - - Updated - - -
That's because EVDO's theoretical maximum downlink speed is about 3.5mbps
.
EVDO Rev A is about 3072 kbps (in 1.25 MHz). The fastest you could see via speedtest would be less than that, of course. Also, I always feel compelled to point out that while support of any individual application (e.g., can I stream HD youtube) is important to the individual user, the way the carriers are looking at it is how many users can they support doing a given task given a single tower.
Consider a hypothetical town being served by a single site (tower) and this is a strange tower with a single geographic sector. One company deploys EVDO-A there, one deploys HSPA+ and one deploys LTE (SISO). The EVDO-A company happens to have 40 MHz available for forward channel (all of them are frequency duplexed), HSPA+ has 20 MHz available and LTE company 5 MHz available. Also presume that streaming HD youtube takes
EVDO company can deliver about 96 mbps, so could support (at 2 mbps per stream) 48 users streaming HD YouTube.
HSPA+ company can deliver about 84 mbps, so could support 42 users streaming HD YouTube at the same time.
LTE company can deliver 27 mbps (really about 22 mbps), so could support 13 users streaming YouTube.
Which is the better technology? It doesn't matter, the company with the most spectrum is winning, even though EVDO-A is using a lower order modulation.
So, let's say you are the HSPA+ company and you want to win. You have 20 MHz of spectrum and you've currently got HSPA+21 deployed, so that means you have 4 carriers each delivering a possible 21 mbps. What does upgrading to DC-HSPA do for you? In terms of the metric above, nothing. The tower is still delivering 84 mbps, it's just that each individual user can now receive up to 42 mbps rather than 21 mbps. What would upgrading to LTE do for you? If you deployed LTE in all 20 MHz you'd have a theoretical bitrate of about 108 mbps, but subtracting coding and control you're down to about 87 mbps - not all that much better than your position with HSPA+21 or DC-HSPA+42, and also you had to change out your equipment at the tower. LTE is really only a little bit better if you keep things "fair" meaning SISO LTE vs SISO HSPA (there are other advantages not associated with theoretical maximum performance, but I think those advantages could be had in more advanced HSPA+ deployments too).
Another option would be to go DC-HSPA+84 MIMO, which would change the metric above so that you're delivering 168 mbps from the tower in 20 MHz and could support 84 users streaming HD YouTube at the same time. This has a lot of nice advantages - voice calls already work over HSPA and you could use every last kHz you have available for HSPA - LTE and EVDO carriers have to keep legacy technologies around to support voice calls at the moment. All of your old HSPA devices will still work, DC-HSPA-MIMO is backward compatible. And, of course, your users can just order a Google Nexus 4 and put it on your network.
It's sort of interesting that T-Mobile went with DC-HSPA rather than HSPA-MIMO, as HSPA-MIMO would have approximately doubled tower throughput while going DC-HSPA doesn't change anything to tower throughput (actually about 10% from improved efficiency), the only advantage is that the end user sees. However, going with MIMO would have required more physical antenna on the tower, and there would be cost associated with that, so that might have been the reason, however if T-Mobile's long term plan were to stick with HSPA there wouldn't be a reason to avoid that upgrade, as the future very-high-speed paths require MIMO anyway.