The iPhone 4S, HSPA+, and When HSPA+ is Real 4G

I've tested a variety of HSPA+ devices on AT&T's networked and never even approached the HSPA theoretical max of 14.4Mbps, let alone HSPA+'s max. More commonly, I'd see speeds of 1-4Mbps down. Never any more than 4.

I think the issue was that Sprint started marketing their WiMAX service as 4G. In real-world testing, Sprint's WiMAX service has consistently scored around the same as AT&T and T-Mobile's networks (http://www.pcworld.com/article/221931/4g_wireless_speed_test..., http://www.pcmag.com/Fastest-Mobile-Networks-2011). Once Sprint had announced that it had 4G and everyone saw that "4G" meant 2-3Mbps, it was hard for AT&T and T-Mobile to avoid calling their HSPA+ networks "4G" as well. After all, they were providing equivalent speeds to Sprint's "4G" network.

The real issue is that there wasn't a good way to label networks that were 2-4x better than their predecessor. EV-DO Rev A achieves around 1Mbps in the real world. As companies (Sprint, really) started rolling out faster wireless technologies, they wanted a way to differentiate them. They were better. If you were a Sprint user with a WiMAX device, you're likely getting 2-3x faster speeds. That's a significant improvement and, frankly, Sprint needed a way to communicate that to users. Of course, that set the bar decently low for 4G - a height that HSPA+ easily surpassed.

It would have been good if it had been called 3.5G or 3.75G, but once Sprint had announced it's 2-3Mbps service as "4G", it meant everyone offering 2-3Mbps service had to call it that. Granted, I think Sprint thought its service was going to provide faster speeds than that - as noted by their 3-6Mbps figure. However, that never panned out in real-world testing. It's reasonable to think that LTE's speeds are going to be reduced as the customer load increases (right now, AT&T and Verizon's LTE networks are virtually empty). So, if WiMAX had been able to deliver 3-6Mbps and LTE speeds went down to the 6Mbps range under load, it might not have seemed so off.

What we really need is to get away from the "G"s and start concentrating on average real-world speed under load. What are people going to call "4G" technologies like LTE Advanced (the next generation of LTE)? It's going to offer vast improvements over the LTE that AT&T and Verizon are rolling out, but officially it's supposed to just be 4G. The number of "G"s doesn't matter. Ping times and bandwidth matter.

Probably a typo... :P

For me the surprise was iPhone 3G back then with Apple skipping the iPhone 2. Apple probably was thinking about iPhone 4G but it wasn't perfect yet or they just wanted to stop marketing hardware features like that.

Other manufacturers have only themselves to blame for that.

Considering that in most parts of the world you don't even get 5GB "unlimited" traffic, considering that real world bandwidth is still ways away from 14Mbit/s and finally considering that even with high bandwidth, latency still is really, really bad, I don't understand this discussion at all.

Even if you somehow reached those 14M consistently, latency would still make sure that surfing felt much slower than over landlines. And the very low data transfer limits will make sure that stuff not as dependent on latency (like video with big buffers) still won't be practical.

For me, the mobile web got kinda usable with 3G phones and since then, I've never seen any practical improvement.

So at least here in Switzerland, this discussion is pointless and it's different features that will make people want (or not - we'll see) to upgrade their phones