Wireless Charging Plans
Wireless operators have recently been plagued by the same type of problem Wireline operators have long experienced with P2P video downloads: 5% of users using 80% of the wireless bandwidth. However, in wireless, the problem is due to smart phone users who run high bandwidth applications like video, mapping, and the like. In addition, developers of attractive apps often use often using multiple flows to increase download speeds. These high capacity users push “normal” users into high loss, low throughput, bad response time service which tends to make them want to switch providers. There are several solutions to this:
- Convert to usage based charging. The challenge is that users typically rebel, preferring a fixed monthly rate with consistent service all month.
- Equalize all users to receive the same peak capacity during congested periods. This leaves the higher capacity users (video, etc) with no viable option. Adding multiple peak rate options at higher prices leads to users finding no differences in service for their high price except during highly congested periods.
- Provide multiple Average Rate priority classes at staggered prices where each class has all active users equalized to the same average rate, including all their flows. This provides the users a clear “all you can use” service which always supports the type of traffic they tend to use. If they watch video all day, they pay more but it works. The difference between classes is clear since when class 1 is getting 100 Kbps average rate, class 2 is getting 500 Kbps average rate, and class 3 is getting 2 Mbps average rate (or some other assortment of average rates and prices).
The key here is to provide different average rates. These will vary by the total demand level, versus different peak rates (which mean nothing in practice) and eliminates the need for the hated usage based charging.
The capability to provide any of these options exists today in many types of traffic management equipment. Anagran (of which, full disclosure, I am CEO), for example, can support all the options. The average rate option is not easy as each flow needs to be rate controlled and is often not available in all equipment. In addition, users cannot switch quickly.
However, I believe offering such a choice will be critical to the future of any wireless data service provider and will ultimately determine which provider grows faster. Though switching services might not be instantaneous, users will do so over time if they are unsatisfied with either the service or the pricing. Such a system provides a fair, time tested solution to the challenge of pricing bandwidth use and encouraging usage growth.
[…] Larry Roberts’ blog post on the wireless billing for one dimension of the […]
Dr. Roberts is the man who designed the ARPANET, the precursor to the Internet that proved packet-switching was a viable concept. We owe him a tremendous debt.
Option 3, multiple average rates, is what Comcast offers. See http://www.comcast.com/Corporate/Learn/HighSpeedInternet/speedcomparison.html where they offer:
Download / Upload
12 Mbps / 2 Mbps
15 Mbps / 3 Mbps
16 Mbps / 2 Mbps
20 Mbps / 4 Mbps
22 Mbps / 5 Mbps
30 Mbps / 7 Mbps
50 Mbps / 10 Mbps
Yes, Comcast is wireline, not wireless, but they have the same issues, in spite of, or perhaps because of, their higher speeds compared to wireless.
Mike, Comcast offers multiple peak rates, not average rates or minimum rates. At peak hours all the subscribers on a given fiber channel (38 Mbps supporting about 400 users) are managed by the peak rate they have elected. However, as the load increases, all flows slow down. The speed tests that are used to prove the speed always work much faster than any other download (easy to arrange). What I meant by average rate was the typical rate the subscriber receives. Perhaps an even better rate to sell is a minimum rate since that can be guaranteed by not over allocating. However, it would typically be much lower than the average rate and users would be put off seeing that.