Venue: Communications and Strategies Journal, no. 63, 3rd quarter 2006, p1
The paper provides a good overview on overlays and attempts to provide a formal definition and taxonomy.
An overlay is a set of servers deployed across the Internet that:
- Provide infrastructure to one or more applications,
- Take responsibility for the forwarding and handling of application data in ways that are different from or in competition with what is part of the basic Internet,
- Can be operated in an organized and coherent way by third parties (which may include collections of end-users).
- peer-to-peer e.g. Napster and Gnutella
- CDN e.g. Akamai
- Routing e.g. RON
- Security e.g. VPNs, Tor, Entropy
- Experimental e.g. PlanetLab, I3
- Other e.g. email, Skype, MBone
The authors assert that overlays do not follow the end-to-end principle because even though from the IP layer's point of view, overlay servers are simply end-nodes, from the application's point of view, they are considered infrastructure.
The paper discusses policy issues and the relationship between industry structure and overlays, asking several thought-provoking questions. It then goes into depth discussing the implications of CDN overlays, security overlays, and routing overlays.
One passage I really enjoyed was the description of why BGP is insufficient:
The paper overall is an easy, entertaining read and gives a nice overview of the issues surrounding overlays and their use and deployment in the Internet.... Broadly speaking, BGP allows each ISP to express its policies for accepting, forwarding, and passing off packets using a variety of control knobs. BGP then performs a distributed computation to determine the "best" path along which packets from each source to each destination should be forwarded.This formulation raises two difficulties, one fundamental and one pragmatic. The first of these is that the notion of "best" is in fact insufficient to fully express the routing task. "Best" is a single dimensional concept, but routing is a multi-dimensional problem. Individual ISPs, in making their routing decisions, may choose to optimize a wide variety of properties. Among these might be 1) the cost of passing on a packet; 2) the distribution of traffic among different physical links within their infrastructure to maximize utilization and minimize congestion - so-called traffic engineering; and 3) performance in some dimension, such as bandwidth available to the traffic or transmission delay across the ISP. Furthermore, because the management of each ISP chooses its own objectives, different ISPs may choose to optimize different quantities, leading to an overall path that captures no simple notion of "best", and rarely if ever is best for the user.A second, pragmatic problem with the current internet routing infrastructure is that it has evolved over time from one in which simple technical objectives dominated to one in which ISPs often wish to express complex policy requirements. For this reason the knobs - the methods available within BGP to control routing choices - have also evolved over time, and are presently somewhat haphazard and baroque. This compounds the fundamental problem by making it harder for ISPs to express precisely the policies they desire, even after those policies are known.