by Andrew Seybold
At the SiRF Location Summit 2.0 held at the St. Regis in San Francisco on September 10, the keynote speaker was Barry West, President of Sprint’s Xohm network. I was to introduce him and as I shook his hand he told me I was causing him some grief with my Commentaries and blogs about WiMAX, Clearwire and Sprint. I told him that all he needed to do was to prove me wrong and I would write an article admitting I was wrong. He went on to deliver a very good keynote that was realistic about WiMAX and Xohm, and I have to admit I was impressed with much of what he said.
Shortly afterward, I received a call from Xohm asking me to come to Baltimore and experience the system prior to its launch on September 29. The press seemed to think the launch would not to be until October 6, but that was because Xohm is planning a coming out party in Baltimore that week.
After my visit to Baltimore, am I ready to write an article that I was wrong? The answer is no. Was I impressed with what I saw and the honesty of the folks rolling out Xohm in Baltimore? Yes, I was. But there are many issues that still need to be resolved in my mind–issues that have more to do with the business model than the technology. So while I was impressed with the network, I am still not convinced there is a business model that can make Sprint and soon the Clearwire/Sprint joint venture any money.
Baltimore will be live by the time you read this. It is scheduled for live service (commercial service) on Monday, September 29. The Xohm folks admit that while they think they are ready they will have to wait and see, and that the system’s first customers will help them shake it down. At launch, this will be a notebook and to the home play only. Handheld devices will come later and the combination voice (Sprint CDMA) and data (WiMAX and CDMA EV-DO) will come later still. But the fact that they are launching a major city is to their credit. Clearwire is still rolling out small cities and not doing well at penetration, perhaps Xohm having a major city under its belt will help it move forward.
The system currently includes 170 cell sites and will by the end of the year be at more than 200 with the ultimate goal of 300 by the end of the first quarter of 2009. Each cell site covers about 1 or 1.5 miles, but on the test drive we also saw cell sites as far away as 3. 5 miles-an amazing difference from the 30 kilometers (18.6 miles) Intel was originally promising. At the moment, coverage is mostly within the city of Baltimore, and our tour took us from the inner harbor around the city and back again. Xohm expects in-building coverage to be to the “second wall” (rooms with windows facing the street will most likely be covered) and that over time it will be offering WiMAX to Wi-Fi devices to cover further into buildings.
Sprint has enough spectrum in Baltimore that its cell sites are using three different channels, each of which is 10 MHz wide. So a cell site with three sectors is making use of 30 MHz of spectrum: 10 MHz for each sector and a reuse of N=3, which means it can reuse each carrier of 10 MHz every three cell sites. It has plans, I am told, to get this down to N=1 (which is what CDMA does today) as the demand for data increases.
The price of wireless cards for laptops will be in the $40-50 range and service pricing will be $10 for a Day pass, $25 for monthly Home Internet service and $30 for monthly On-the-go service. Sprint will also offer contracts for one and two years that will be discounted, but was not able to share the pricing with me prior to the launch. It is heading toward one person or one family pricing, so multiple devices put on the network by a single individual will be included in the pricing model and there are safeguards in place to ensure that customers don’t buy multiple devices and simply hand them out to their friends.
During the course of the conversation, the Xohm folks mentioned that many of the sites will be using microwave and/or fiber for backhaul but some will be using multiple T1 lines that will be a choke point until they can replace them with fiber or microwave links with more capacity. When asked how many more cell sites they needed at 2.5 GHz to cover the city than they needed at 1900 MHz for their CDMA system, they said that at the moment it is only about 15% more sites. I have to admit that this number surprised me because of the difference in frequencies (higher frequency means less coverage), but it was clear that in using this formula they anticipate some speed issues at the edges of their cell sites-as is true with all broadband systems.
Xohm intends to “promise” its customers download speeds of between 2 and 4 Mbps and upload speeds of a megabit or more. During the tests driving around the city, we experienced data rates from 6 Mbps to under 1 Mbps and upload speeds from under 500 Kbps to over 1.4 Kbps. Keep in mind that this was on an unloaded network that had not been turned on for commercial service and the tech types were still tuning it for maximum performance.
During the drive tests I saw multiple video feeds, heard VoIP that sounded pretty good and saw real-time web surfing. I also saw several disconnects, several times when the speeds were less than what they want to deliver and a few bugs in both the software and the network. All of this, they assured me, would be taken care of before the launch. I was impressed with the performance of the network overall, realizing, of course, that it was not live and that there was virtually no loading at this point. I was reminded of my time in the Qualcomm van, running around San Diego testing both LTE and UMB, and the results from an end user point of view appear to be about the same.
Sprint is making a case that it has a lot more capacity than LTE will, simply because of the bandwidth it has available, and also made the point that it is not competing with EV-DO or UMTS/HSPA, but rather wants customers who are in the 4 GB or more data usage category. Sprint feels that its pricing will be more than competitive and doesn’t believe the AT&Ts or Verizons of the world want this type of data customer on their networks.
As I see it, we now have four types of “hotspots” for consumers and business customers to consider. The first is Starbucks and other Wi-Fi hotspots, the second is Sprint Xohm where your city is your hotspot, the third is Verizon and AT&T with the United States as your hotspot and the fourth is Qualcomm Gobi chipset groups where the world is your hotspot. The question that remains unanswered is which type of hotspot consumers will pay for and how much.
When asked about regional or citywide coverage vs. nationwide coverage, the Xohm folks were quick to point out that only a small portion of customers travel beyond their own city or region and, since the majority of customers tend to stay put, this system will be appealing to them. They also made it clear that they do not believe they are competing against EV-DO (Sprint or Verizon) or HSPA (AT&T and T-Mobile) but are going after customers 3G providers don’t want. They did try to convince me that WiMAX was a 4G technology, but frankly I am not buying it. There is no definition of 4G and if you run a comparison of data throughput based on the same amount of spectrum, you end up with about the same bits/Hz ratio for all of the 3G services. Xohm has more spectrum, therefore it has more throughput available, but it also pays the price of being in the 2.5-GHz band where it needs more cell sites per city, even if Intel says differently.
The bottom line is that from what I saw of the demo and my discussions with not only those in marketing, but with the engineers building the network, the WiMAX technology works-which I have never doubted. What remains to be seen is if the American thirst for the Internet on the go will provide Sprint’s Xohm with enough customers to pay the bills and turn it into a profitable venture. It is also unclear what the merger with Clearwire will do to the business model, and really unclear if a data-only network, with VoIP at some point, can compete with existing networks that offer voice and data services.
It is easy to be impressed when you are sitting in a van that is equipped with all of the best technologies available and the route has been pretested over and over again. The tests were impressive; the dropped calls were not minor but, it was pointed out, this is still a work in progress as every new system once was. I was impressed, but I will be more impressed if Xohm or Clearwire has an attach rate, ARPU rate and churn rate that suggest success a year from now. They have already set up the daily customer standard so the standard wireless matrix may not be the best way to measure their success. Whatever the measure, their success will be determined a customer at a time and the real test will be how many daily customers become yearly subscribers. Time will tell, and I wonder what throughput numbers I will see if I take the same test drive again in a year. The wired Internet folks are just now finding out that bandwidth, even wired bandwidth, is not unlimited. I wonder how long it will be until Xohm/Clearwire are not so keen on customers who use 4 or more gigabytes of data per month.