BITAG Report on Traffic Differentiation

The Broadband Internet Technical Advisory Group (BITAG) is a non-profit, non-partisan group of technical experts from the communications, content, and public interest communities that examines issues related to Internet regulation and makes recommendations. Quoting from its web site, BITAG was born out of the realization that “a broad cross-section of the Internet community saw the need for, and value of, a technical advisory group to discuss and opine on technical issues pertaining to the operation of the Internet, as a means of bringing transparency and clarity to network management processes as well as the interaction among networks, applications, devices and content.” BITAG has issued eight technical reports on Internet management so far.

The most recent BITAG report deals with Internet traffic management. It’s essential reading for anyone who wants a better understanding of how the Internet works and why it works that way; I’m biased as I’m a member, but I think readers of the report will agree.

The traffic management report makes six key observations:

  1. TCP causes recurring momentary congestion.
    • When TCP transfers a large file, such as video content or a large web page, it practically guarantees that it will create recurring momentary congestion at some point in its network path. This effect exists by design, and it cannot necessarily be eliminated by increasing capacity.  Given the same traffic load, however, the severity of the momentary congestion should decrease with increased capacity.
  2. A nominal level of packet discard is normal.
    • Packet discard occurs by design in the Internet. Protocols such as TCP use packet discard as a means of detecting congestion, responding by reducing the amount of data outstanding and with it self-induced congestion on the transmission path. Rather than being an impairment, packet discard serves as an important signaling mechanism that keeps congestion in check.
  3. The absence of differentiation does not imply comparable behavior among applications.
    • In the absence of differentiation, the underlying protocols used on the Internet do not necessarily give each application comparable bandwidth. For example:
      • TCP tends to share available capacity (although not necessarily equally) between competing connections. However, some applications use many connections at once while other applications only use one connection.
      • Some applications using RTP/UDP or other transport protocols balance transmission rate against experienced loss and latency, reducing the capacity available to competing applications.
  4. Differentiated treatment can produce a net improvement in Quality of Experience (QoE).
    • When differentiated treatment is applied with an awareness of the requirements for different types of traffic, it becomes possible to create a benefit without an offsetting loss. For example, some differentiation techniques improve the performance or quality of experience (QoE) for particular applications or classes of applications without negatively impacting the QoE for other applications or classes of applications. The use and development of these techniques has value. 
  5. Access technologies differ in their capabilities and characteristics.
    • Specific architectures and access technologies have unique characteristics which are addressed using different techniques for differentiated treatment.
  6. Security of traffic has at times been downgraded to facilitate differentiation techniques.
    • Encrypted traffic is on the rise and it has implications for current differentiation techniques. In response to this increase, some satellite and in-flight network operators have deployed differentiation mechanisms that downgrade security properties of some connections to accomplish differentiation. The resulting risks to the security and privacy of end users can be significant, and differentiation via observable information such as ports and traffic heuristics is more compatible with security.  

Of these observations, the ones I find most compelling are 3, 4, and 6. Internet policy advocates of various stripes – from content aggregators to application developers and public interest advocates – have long insisted that ISPs should treat all traffic indiscriminately, treating every packet the same as every other packet. But BITAG does an analysis of the effects that policy choice has on commonly used applications and finds them undesirable. This goes to observation 3, The absence of differentiation does not imply comparable behavior among applications.

One simple case addresses what happens when VoIP is running on a last-mile broadband pipe at the same time as video streaming. Although video streams tend to be rate-limited by the sender at a more or less constant rate over minutes, this average is the function of alternating between furious periods of transmission punctuated by long periods of silence. In other words, the video streamer sends hundreds of packets back-to-back as fast as it can for some period of time, and then goes dark for a period of several seconds.

This cyclic form of transmission means that the last mile connection will effectively alternate between congestion and silence. This pattern co-exists well with other video streaming applications and reasonably well with web surfing, because these applications exhibit a similar pattern of overload and underload. But it’s deadly for VoIP and video conferencing because they exhibit a different pattern: they send and receive equal amounts of traffic at a regular interval. VoIP also depends on the speedy delivery of its packets while video streaming does not.

Streaming packets can be re-ordered without affecting their average transmission rate if the ISP simply “paces” them at the average rate, essentially converting a clumps of data into streams with more regular spacing. When video streaming is de-clumped, VoIP and conferencing work better and streaming works as well as it does when it’s not paced. Hence, observation 4, Differentiated treatment can produce a net improvement in Quality of Experience (QoE).

This is obviously important today because video streaming represents half the traffic on the North American Internet during prime time and lots of people use Skype, Vonage, and similar services.

Observation 6 addresses practices employed by inflight Internet services that compromise security in order to achieve a crude form of network management. BITAG doesn’t like that and would like to see these networks managed better.

While it’s undeniably good for ISPs to de-clump video streams, BITAG recognizes that there are trust issues with some ISPs and some forms of traffic management. Hence, its recommendations emphasize disclosure and management that improves network quality of experience:

  1. Network operators should disclose information on differential treatment of traffic.
    • In previous reports, BITAG has recommended transparency with respect to a number of aspects of network management.  BITAG continues to recommend transparency when it comes to the practices used to implement the differential treatment of Internet traffic.Specifically with respect to consumer-facing services such as mass-market Internet access, network operators should disclose the use of traffic differentiation practices that impact an end user’s Internet access service. The disclosure should be readily accessible to the public (e.g. via a webpage) and describe the practice with its impact to end users and expected benefits in terms meaningful to end users. The disclosure should include any differentiation amongst Internet traffic and should disclose the extent and manner in which other services offered over the same end user access facilities (for example video services) may affect the performance of the Internet access service.
  2. Network operators and ASPs should be encouraged to implement efficient and adaptive network resource management practices.
    • In a previous report BITAG recommended that ASPs and CDNs implement efficient and adaptive network resource management practices; we reiterate that recommendation here, extending it to network operators. Examples of such practices might target the minimization of latency and variation in latency induced in network equipment, ensuring sufficient bandwidth for expected traffic loads, and the use of queue management techniques to manage resource contention issues.
  3. Quality of Service metrics should be interpreted in the context of Quality of Experience.
    • Common Quality of Service metrics, often included in commercial service level agreements, include capacity, delay, delay variation, and loss rate, among other things. From the viewpoint of the end user application, these metrics trade off against each other and must be considered in the context of Quality of Experience. For example, since TCP Congestion Control and adaptive codecs depend on loss to infer network behavior, actively trying to reduce loss to zero leads to unintended consequences. On the other hand, non-negligible loss rates often directly reduce the user’s Quality of Experience. Hence, such metrics should be interpreted in the context of improving user experience.
  4. Network operators should not downgrade, interfere with, or block user-selected security in order to apply differentiated treatment.
    • Network operators should refrain from preventing users from applying over-the-top encryption or other security mechanisms without user knowledge and consent. Networks should not interfere with, modify, or drop security parameters requested by an endpoint to apply differentiated treatment. Given the potential for possible exposure of sensitive, confidential, and proprietary information, prior notice should be given to end users of traffic differentiation features that affect security properties transmitted by endpoints.

Among recommendations, numbers 2 and 3 are most important. BITAG has previously recommended that ASPs and CDNs implement “efficient and adaptive network resource management practices” and it now extends this recommendations to ISPs themselves. The example of de-clumping video streams makes the case for this recommendation all by itself, but the report offers a number of other examples that reinforce it.

The end goal of traffic management practices (including adding super bandwidth in gigabit networks) is to improve overall Quality of Experience as perceived by users of diverse applications. Simply adding capacity does not prevent streams from fighting each other, as streaming does to VoIP, it simply reduces the number of perceived VoIP failures by a small amount. Pacing is much more effective, even though there are good reasons to increase capacity as well.

The major insight in this report is the need to judge management practices by their subjective effect on end-users running Internet applications. There was very little thought given in the development of Internet standards and practices to how things like TCP congestion control affects other applications, hence reconciling TCP’s peculiar operation with the needs of VoIP users falls on ISPs.

BITAG wants ISPs to play an active role, but to do so sensibly; the way the satellite operators downgrade security is not sensible management. Hence, BITAG stresses that Network operators should not downgrade, interfere with, or block user-selected security in order to apply differentiated treatment.

BITAG also counsels against inter-application management that reduces the average rate of any application, but you’ll have to read the report or wait on my next BITAG post for the details.