Quality of service

National regulatory authorities (NRAs) and policy makers are concerned to ensure that customers receive optimal levels of communications service. There are many aspects to ‘quality’, and NRAs across the world have adopted a variety of approaches to encouraging or requiring higher quality.

A selection of these aspects of quality is briefly discussed below, together with approaches adopted by regulators in these cases.

Quality of Service

Is the geographical coverage sufficient?

For fixed voice services in the developed world, this is ‘universal’, often originally provided under a state monopoly and then maintained by a universal service obligation (USO) imposed by regulation on a universal service provider. For mobile services, mobile coverage has historically been well provided by the market, though customers can be encouraged by NRAs to check competing offers by using tools such as coverage checkers. Mobile and broadband services are perceived as becoming more essential to society, and extending coverage beyond the commercially rational limits is now a political issue in many countries. Specific gap-funding interventions (compatible State Aid) are being deployed in some cases; in other cases regulatory mechanisms are being used such as coverage obligations in spectrum licences or (much less frequently) mandating operators to offer national roaming. Encouraging different types of site or network sharing while maintaining incentives for competition can also improve network coverage by reducing the costs of providing rural coverage.

Is there enough capacity?

In the voice world, this is known as ‘grade of service’. The mathematics needed to engineer an adequate voice solution for most practical purposes was provided by Erlang roughly a hundred years ago. Prioritised data systems can be designed and built to provide guaranteed capacity, but for Internet data we are still largely using best-efforts mechanisms rather than prioritisation, limiting the usefulness of the Internet for some service types. Inter-provider links using prioritisation are uncommon, but this has not stopped the European Commission from making regulatory proposals for ‘Assured Service Quality’.

Does the service available at your location have the required capabilities?

For data services, this is not a simple question. Does your broadband line provide enough capacity in the busy hour? What about peak speed, latency, jitter, packet loss? Beyond such technical performance parameters, what about the end-user applications’ performance (e.g. could the broadband service be used to provide voice services ?) Sometimes retail competition is not sufficient to ensure that end users’ needs are met. For example, in New Zealand, the NZCC consulted on the non-price parameters of the regulated bitstream service (UBA) provided by the structurally separated operator Chorus (which has no retail customers). This topic also has links to broadband USO, USO (through voice over broadband), and to net neutrality especially if the ISP is also providing other services over the same links or if the quality provided varies with the IP address, protocol or port.

Are local net neutrality rules being respected?

Various countries have instituted specific net neutrality regimes which prevent ISPs from discriminating between packets for different applications and services. The EU Regulation also places restrictions on the ability of ISPs to offer “specialised services” (which are services other than Internet access services which are optimised for specific content, applications or services, or a combination thereof). Regulators have a role in monitoring whether these rules are being respected.

For further discussion on this topic, see Having your cake and eating it: network slicing and net neutrality rules in Europe.

Is service interrupted too much by faults? How long should they take to fix? How long should you have to wait to get service? How should this be specified? What is the relationship between operating costs and time to fix faults?

A monopoly supplier with regulated prices might have incentives to underinvest in reliability or in its ability to fix faults rapidly. In some cases regulators have specified a fault rate key performance indicator (KPI) as part of the USO, with material financial penalties if these targets are not met.1 For mobile networks, local faults are often not noticed by end users, but if central equipment faults do occur, these can deprive millions of people of service for many hours.

In other cases, operators and regulators have modelled the impact of various system parameters (notably, the number of staff) on the SLA performance statistics, to investigate the potential tradeoff between cost and quality. The figure below is one from an Ofcom model built in conjunction with Analysys Mason, and shows the way in which (in a certain modelled scenario) the staffing levels influence the percentage of the time that the different services hit their respective service levels. The performance is strongly non-linear: if faults arise faster than they can be fixed on average, then the queue grows and grows, and the performance is very poor; at a certain level of resources most faults can be fixed in a timely way; and then a point on the S-curve is reached at which increasing performance takes significant additional quantities of resource. Finally, other factors can place a maximum limit on the performance that can be provided (e.g. there could be a small fraction of faults that cannot be remedied in the time available irrespective of the resource levels).

Figure 1: Illustrative output from the simulation [Source: Analysys Mason (using Ofcom RPM), 2017]

Illustrative output from the simulation

The model used to generate this chart is a simulation built in Python, which is described here.

Alternative techniques may exist for some simpler situations, such as the use of queueing theory (e.g. M/M/c queues) which (if sufficiently similar to the problem at hand) can give more readily understandable insights– which takes us back to Erlang’s work.

Is the advertising misleading?

Advertising regulators and telecoms regulators have both been involved in setting standards in particular around marketing broadband services (especially in relation to speeds offered2) and price comparisons.

1 For example, see http://www.comreg.ie/_fileupload/publications/ComReg14129.pdf

2 https://www.asa.org.uk/News-resources/Media-Centre/2016/ASA-signals-need-for-change-in-advertising-of-broadband-prices.aspx#.Vw5es57mqow

How we can help you

Examples of recent projects in this area completed by our experts include:

  • Analysys Mason undertook a study for the European Commission to develop a guide to investing in broadband network infrastructure. The aim of the guide was to document best practice in combining public and private investment to deploy next-generation broadband networks, and so help EU Member States to absorb available funding and move towards meeting the Digital Agenda for Europe targets. We conducted interviews with 12 existing broadband projects, and combined this with our own expertise to develop a step-by-step guide to planning a broadband investment. The guide provides advice for small regional roll-outs and major national interventions alike.
  • A Western European regulator engaged Analysys Mason to assess the value of i) several spectrum bands currently used in the client’s country for which the licences were due to expire in the next few years, and ii) bands that were not currently allocated but might be in the future. As part of the project, we also conducted a detailed benchmarking exercise on the various spectrum award mechanisms used internationally (auctions, beauty contests, etc.), the regulatory conditions imposed on licences (coverage obligations, number of blocks and amount of spectrum (MHz) per licence, etc.), and the reserve prices set for the existing and new frequency bands.
  • We assisted an EU NRA in the implementation of a set of market reviews under the European Regulatory Framework, in collaboration with a legal partner. This included the relevant product markets for broadband (including wholesale and retail broadband markets), with particular emphasis on issues such as next-generation access and competition between DSL and cable. The aim of the project was to enable our client to: prepare a definition of each relevant market, analyse the level of competition, and identify remedies to be applied to each market to promote competition.
  • For BEREC, we studied the approach taken by NRAs in selected non-European benchmark countries (Chile, India and USA) to address net-neutrality issues, including the approach taken to monitoring QoS.
  • We recently ran a workshop for an EU NRA focussed on its duties under the new EU Regulation covering net neutrality.
  • As part of Ofcom’s previous WLA market review, Analysys Mason reviewed a model provided by BT which related the service level (in terms of the percentage of faults not fixed and lines not installed within the required period) to the staffing levels in the field force.
  • We subsequently reviewed a new BT model of the QoS impact of different resource levels. Our report was published by Ofcom here.
  • Analysys Mason conducted an independent review of New Zealand Telecom’s mobile network after two major incidents which denied service to significant numbers of customers in 2009 and 2010.
  • We conduct audits of price comparison sites under a price comparison accreditation scheme.

Featured Articles More articles