Four Carriers, Four Ideologies: Assessing The Cellular Data Strategies
Greetings from Orlando. As I sit here fighting with molasses-slow hotel “broadband”, AT&T cellular service that rapidly alternates between four bars of 3G, four bars of EDGE, and ‘Searching…’ (and, even when it reports solid service, often won’t serve me any data…remind me why I switched, again?), and Verizon cellular service that isn’t much better, my thoughts wander to the short- and long-term plans of the four major U.S. cellular providers. It’s a topic I most recently touched on in a post-India-trip early-February writeup.
It’s a timely topic, given that both early yesterday morning (via Lisa Su, Ph.D, Senior Vice President and General Manager, Networking and Multimedia) and late yesterday afternoon (from Brett Butler, Vice President and General Manager, Network Processor Division) I heard about Freescale’s sustained, strong presence in and continued dedication to the cellular base station business. And it’s a dedication that Freescale backed up with two key announcements this week; a quad-core 32-bit Power family and its first 64-bit Power CPUs. Along with learning about amazing new technologies and products and interacting with (and learning from) really smart people, having in-depth access to and being able to analyze business strategies of the various companies I cover is one of my favorite aspects of this job.
The second-smallest carrier of the U.S. ‘big four’ as measured by subscriber counts, Sprint, is the most aggressive from a next-generation technology rollout standpoint. Via its majority ownership of Clearwire, the company is in the midst of an aggressive rollout of 4G WiMAX technology across the United States. CLEAR WiMAX service is currently in several dozen markets, and the company plans to bring CLEAR to 80 markets covering up to 120 million people by the end of 2010. To wit, signs of network life are now beginning to appear in all-important Los Angeles, New York City and San Francisco. Sprint offers 4G-only and more flexible mixed 3G/4G USB modems for computers, along with cellular hotspots. Mini-PCI Express add-in modules are even on the way. And the company recently released its first WiMAX data-supportive handset in partnership with HTC. The EVO 4G, an Android-based device, was panned for poor battery life, but that’s typical for a first-generation design. All in all, though, reviews were uniformly quite positive, the handset is selling well, and WiMAX follow-ons from Samsung and other companies are on the way. Voice traffic currently still routes over CDMA, although migration to a VoIP-like voice-over-WiMAX approach is planned.
The fundamental reason for Sprint’s WiMAX rollout aggressiveness is alternative 4G cellular technology LTE, specifically as planned by historical CDMA competitor Verizon (the current #1 U.S. carrier as measured by number of subscribers). Verizon’s made no bones about its intention to be first to market across the U.S. with worldwide industry standard LTE. In her presentation yesterday, Dr. Su accurately pointed out that only focused-market technology trials are underway right now; even though handsets are forecasted to soon appear (up to five by the middle of next year), she doesn’t expect LTE service to roll out in earnest until 2012 (or later). Nonetheless, operators worldwide are buying equipment now, which they expect will be upgradeable to support not-yet-finalized capabilities such as voice-over-LTE (not to mention voice-over-Revision A). As such, Su and her lieutenants at the breakfast press briefing pointed out the flexibility afforded by software-powered DSPs versus more hardwired solutions. It’s a pitch that I’ve also heard innumerable times from the FPGA providers in contrasting their products versus ASICs, and it’s a pitch I’ve also successfully used myself innumerable times in my past life as a flash memory advocate at Intel. And speaking of flexibility, both Clearwire and Sprint have publicly admitted that if WiMAX momentum falters, they won’t hesitate to migrate to LTE, which overlaps in some key feature compatibility areas (or not, depending on who you talk to).
#2 AT&T is, as most of you are probably already aware even if you aren’t one of its cellular customers, widely panned for its network’s poor voice and data service. The carrier claims that it’s making ongoing infrastructure build-out investments (supplemented by its customers’ own broadband connections) and even offers an iPhone application that lets you report the locations where you’ve experienced problems (after those problems disappear and your connection is resurrected, of course). And to its (minor) defense, AT&T is the sole U.S. carrier for the iPhone, which has proven to be the most successful (therefore the most data-hungry) smartphone to date, with various carriers’ Android-based designs in second place and steadily catching up. AT&T also plans to implement LTE, although it’s generally believed that the carrier is a year or more behind Verizon. In the interim, AT&T is moving to migrate many of its markets to speedier 7.2 Mbps-and-beyond UMTS HSPA flavors. The fundamental hinderances to the carrier’s aspirations, as my experience of the first paragraph exemplifies, are its ‘backhaul’ connections between the cellular base stations and the remainder of its wired backbone network infrastructure. There’s no value, only frustration, associated with a strong handset connection to a cell tower if the voice and data traffic is choked to the point of impassability beyond that point.
And then there’s #4 carrier T-Mobile. I’m finding it a little ironic to even write these words, since regular readers know that the carrier is tower-deficient in my rural home office locale and also still relies on archaic GPRS as its cellular data technology there. Yet, in more population-rich areas, T-Mobile plans in the short term to one-up both AT&T and its EV-DO-based CDMA competitors with HSPA+, which touts downstream data rates as high as 21 Mbps. The carrier hasn’t yet made public its (true) 4G plans, whether they be LTE or WiMAX, and I agree with AT&T’s complaints that T-Mobile’s trying to misrepresent HSPA+ as ‘4G’ in the absence of an announced next-generation approach. T-Mobile’s task is also complicated by the fact that it doesn’t use the same 3G frequency band as AT&T and, therefore, requires custom handsets that support 1900 MHz. But given that T-Mobile’s small size limits its investment capabilities, a HSPA+ upgrade is a cost-effective alternative strategy…especially given that the carrier is believed to have far more robust ‘backhaul’ capabilities than AT&T, therefore providing higher likelihood that the 21 Mbps potential will translate into some reasonable semblance of reality.
Readers, which of these strategies do you think will pan out, both in the short- and long-term?