Friday, September 25, 2009

Net Neutrality: Do You Want Random or Planned Blocking?

Strong forms of "network neutrality" that allow no prioritizing of Internet traffic pose clear issues for real-time services such as video and voice, at times of peak load. Since all real-time services are highly susceptible to congestion, peak hours may not be optimal times to use them. The problem is that "peak hours" are when users most want those services available.

So the issue is simply whether users prefer random disruptions of quality, or planned disruptions of the quality of some applications, in order to provider optimal performance for real-time services. Since no federal rules are going to abolish peak consumption hours, the only practical issue is what methods are used to limit network access at times of peak congestion.

One can specify that no packets are blocked or intentionally slowed. But that only shifts the congestion control mechanism. Users will find their voice and video sessions don't work that well, and will restrict their use of those services at times of peak congestion.

The other approach is to prioritize applications so the burden of congestion is shifted to lower-priority communications or applications; to users with demonstated "very-high usage" profiles; users who already have exceeded fair usage caps or users at congested cell towers.

Some service quality degradation, at peak hours, will happen. The only issue is how the congestion issues are distirbuted.

One might argue that congestion effects also could be distributed based on whether users have opted to buy "best effort" usage plans or "assured access" plans. But higher-quality, assured access is not possible when packets cannot be prioritized.

Congestion issues, despite network upgrades, always will be a fact of life for networks. The only practical issue is how network degradation is handled. It can be random, or it can be planned. How any new network neutrality rules are framed will determine whether planned mechanisms are allowable.

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...