Load balancing

A couple of weeks ago, I wrote about the artificial distinction that is being maintained between business and residential lines on a wholesale level.

In the olden days, these distinctions made sense. We cross subsidized to keep residential rates low, justifying it based on the value of service and all sorts of reasons that made sense in an era of regulated rates. In the case of internet service, especially wholesale internet service, this distinction no longer makes sense.

In many cases, residential users are more demanding on their internet service provider than their business clientele. Imagine an airline that insisted that all people travelling for business purposes had to buy business class seats and no vacationers were ever allowed up front. Or if businesses had to buy different printer cartridges and paper than school kids – by government fiat. There may be very good reasons to buy a different grade of office supplies, but it is up to the purchaser to decide.

I continue to be troubled by the resolution of the wholesale internet access decision. There are many who think the rate setting process was unfair because competitors aren’t given the chance to go through the incumbents’ financial details. They need to get over that; we have competent regulators who are able to review the numbers, balancing the public interest with reasonable provisions of corporate confidentiality.

The wholesale regime correctly tries to emulate the way that networks are built – with lumpy amounts of capacity being acquired by the independent ISPs and then managed as best they can toward profitability.

Except for one big management tool. The way the wholesale regime got structured, independent ISPs can’t mix business and residential customers on their access network connections because the CRTC created two different wholesale regimes.

Since the beginning of time – at least since the beginning of telecom networks – service providers have been able to take advantage of non-coincident busy periods. Residential customers make greater use of telecommunications in the evenings and weekends; businesses place heavier demands during weekdays. It’s why we call them business hours.

So when intercity and interconnecting facilities get built, they are engineered for the busiest period of the week and the rest rides for free.

The capacity-based billing for ISPs was the intent of the CRTC approach. It is sound policy except for one important piece. By artificially separating business traffic from residential traffic, independent ISPs have to buy double the real requirement. They need to buy the capacity for their business peak separate from capacity required for their residential peak. In a normal network world, the traffic would be intermixed and there would be just one peak driving the capacity requirements for the interconnecting group.

There are strong public policy reasons to combine business and residential traffic: it will increases overall competitive market intensity. Allowing the mixing will ensure that ISPs try to balance their business users with residential clients, with an eye toward keeping traffic loads balanced.

The independent ISPs have expressed concerns about the rates that were set in the CRTC’s UBB decisions [business, residential]. I am somewhat inclined to think that the rates may be close to the right numbers, but the problem is more fundamental. The inability to combine residential and business traffic means that independent ISPs will be unable to optimize the utilization of the facilities through balancing of their customer mix.

It is a point worth re-examination.