The big telecom and cable behemoths would love to charge high bandwidth services extra for fast service. One variation of this proposal would offer these services under a tiered rate structure where you would pay more money for preferential bandwidth. That's what Shaw Communications wants to do.
Despite the cries of many (including yours truly) that this idea cannot be justified, it is not that we haven't seen this concept before. In fact, there have been tiers since there has been an Internet. Now, it's just that companies such as Shaw and AT&T are saying this out loud.
A tiered Internet structure has been with us since the early days of an MCI/academic administered Internet," writes Anthony Mitchell in an Always-On Network Internet post entitled Tiered Internet Already Exists.
"Major ISPs are already instituting preferential handling for different types and sources of traffic. QoS variations already exist for prioritization of certain types of packets. Streaming video is sent platinum, for example," he correctly writes. "Email is bronze. "
Mitchell then describes how this plays out in terms of what he believes is the practice of some providers of both business and residential DSL to prioritize business data packets. The result for consumers, Mitchell believes, is that non-prioritized packets can be delayed or dropped.
The core issue for VoIP carried on these networks, is that there is little or no margin for error. Packet deprioritization thus causes effects that are immediatley noticeable.
"VoIP providers such as Vonage and Packet8 need actual 128 kbps upstream speeds for optimal quality, 80 kbps for decent voice quality. A lot of residential DSL service being sold today in the U.S. does not provide that level of service" he writes. " It is difficult and frequently impossible to provide QoS in home routers. Residential service is often not sold with QoS guarantees.
"Firms such as Internap are providing traffic optimization services by leasing redundant backbone capacity and using border gateway protocol to route urgent traffic on the fastest routes," he notes. "But QoS on the network edge has yet to be systematically addressed, especially for QoS-sensitive applications such as VoIP."
There are great startup opportunities here