Column
EV-DO Rev A: Upping the Ante
Network Computing Mobile Observer, February 7, 2007
By Peter Rysavy
Both Sprint/Nextel and Verizon Wireless have just announced the availability of their Rev A versions of CDMA2000 EV-DO technology, which significantly boosts uplink speeds. This technology is the logical next step in the evolution of wide-area wireless.
On Jan. 30, Sprint announced it's CDMA2000 EV-DO Rev A (Code Division Multiple Access 2000 Evolution Data-Optimized Revision A) network as now available in south Florida; Portland, Oregon; and Puerto Rico, adding these areas to coverage that encompasses 95 million people in 21 other major markets. Sprint calls the service its "upgraded Mobile Broadband Network." Rev A is an upgrade to Rev 0, which became available in 2004, and ups the ante in wide-area wireless by presenting the highest performance claims yet for any available wide area wireless technology.
Specifically, Sprint/Nextel indicates a significant boost in typical uplink speeds that increases previous EV-DO Rev 0 speeds of 50 Kbps to 70 Kbps to a range with Rev A of 350 Kbps to 500 Kbps. Improved uplink performance is no surprise as it has been on technology road maps, including ones I have published, for many years. What is a surprise to many in the industry, though, are greatly increased throughput speed claims on the downlink. Sprint/Nextel originally projected 450 Kbps to 800 Kbps for Rev A, but in last week's announcement the company indicated average transmissions between 600 Kbps and 1.4 Mbps. When a Network Computing colleague asked Sprint/Nextel about this, the response was that these faster speeds more accurately portray the range of speeds customers are likely to experience, based on Sprint/Nextel's actual performance tracking.
These values are quite a bit higher than what one would expect based on the increased spectral efficiency of Rev A over Rev 0, so if Sprint/Nextel can deliver these higher speeds, it may have more to do with improved transport technologies between the base station and the core network, rather than the radio technology.
Not to be outdone, on Feb 2, Verizon Wireless announced its Rev A network as being available in the greater Chicago area. I anticipate Verizon Wireless will upgrade the rest of its network this year. Its throughput claims, however, are much more modest, at 450 Kbps to 800 Kbps on the downlink and 300 Kbps to 400 Kbps on the uplink. Cingular Wireless, with its HSDPA (High Speed Downlink Packet Access) network,0 quotes 400 Kbps to 700 Kbps typical download speeds
with bursts to 1 Mbps or higher. HSDPA and EV-DO are both highly optimized CDMA technologies employing many of the same radio mechanisms, so at the end of the day, I would expect all these technologies to provide similar performance. Keep in mind that there are no rigorous definitions for "average user throughput," as many variables, including number of users in the cell sector (both voice and data), types of applications being used and average user signal quality,
are involved. So all these numbers are somewhat slippery.
It's important to note that enhancements for wireless technologies come bundled in major releases that necessitate network upgrades and generally new devices. In this respect, there has been a long-term leapfrog game going on between the GSM/UMTS (Global System for Mobile Communications/Universal Mobile Telecommunications System) and the CDMA2000 community. It began with GPRS (General Packet Radio Service) for GSM, which was top dog until CDMA2000
1xRTT (One Carrier Radio Transmission Technology) came along. Then Enhanced Data Rates for GSM Evolution (EDGE) jumped past 1xRTT. Then EV-DO Rev 0 leaped forward, followed last year by HSDPA. Now Rev A makes claim to being the fastest technology. However, HSDPA this year will be enhanced with a technology called HSUPA (High Speed Uplink Packet Access), and HSDPA
combined with HSUPA will be called simply High Speed Packet Access (HSPA). HSPA
should match Rev A in both downlink and uplink capabilities.
Enough about competition and claims. There are some other important aspects to Rev A worth understanding. One is that with a faster uplink, combined with QoS mechanisms in the network, it's feasible for carrier-provided VoIP. This is on the road map for CDMA2000 operators, but due to the complexity involved, I expect it's still a couple of years out. Rev A also reduces latency from
the typical 250-msec value for Rev 0 that I have measured, but I haven't seen any numbers published yet. This will help applications, especially chatty ones, as much as increased throughputs.
Rev A devices will include PC Card modems, USB modems and embedded Express card modules. Usage pricing has not changed. Sprint is charging $59.99 per month for a two-year contract, and a usage based plan of $39.99 for 40 MB, which is consistent with other operators. These prices still are a major obstacle for broader adoption, in my opinion, but more users at lower price points clogging up the network are apparently not in operator best interests. That's supposedly where WiMAX will come in, letting Sprint/Nextel leverage spectrum holdings much greater than it has for CDMA2000 to make wireless broadband available at lower price points.
Bottom line, I'm excited by the continual improvements in wide-area wireless technology, and Rev A is an essential step to a ubiquitous mobile broadband future.
Return to column index |
