Much has changed in the 50 years since the invention of packet switching and the early network designs and deployments that would evolve into today's Internet. Network size has grown from a number of nodes that could be counted on our hands to a number estimated to be upward of 800 million network-attached devices and an equally large number of intermittently connected mobile devices. We have also witnessed the transition to commercial Internet use in the 1990s; the creation of the World Wide Web, P2P networks, Voice-over-IP, and streaming media services; and the increasing importance of security and privacy.
And yet, amid this change, the Internet's architecture and a number of the key protocols that form its technical foundation were conceived and subsequently standardized more than 30 years ago. Certainly, there have been changes and improvements in the intervening years, but it's remarkable indeed that the designs that began with Cerf and Kahn's original 1974 landmark paper1 have proven so durable in the face of so much changea tribute to the prescience and wisdom of the Internet's original designers.
While early computer networks focused on "delivering data between computers or between computers and terminals,"1 today's Internet is arguably more concerned with connecting people with content and information. Cisco's Visual Networking Index noted that P2P and video traffic together account for nearly 80% of consumer Internet traffic in 2011. In such a content-centric worldview, what a person wants, rather than where it is located, is what matters most; content, rather than the server on which content resides, becomes the starting point. A number of the current networked applications, including content distribution networks (CDNs), publication-subscribe (pub/sub) systems, and P2P file-sharing applications, are already content-focused. But they do so as applications running on top of the Internet and (importantly) there are many separate CDNs, pub/sub systems, and P2P file-sharing services.
With an increased focus on content and with nearly 50 years of networking research to draw on, the time may be right for a new network architectureone in which content distribution and retrieval, rather than host-to-host packet delivery, is the core function supported in each and every network node. This new architecture might literally be built upon the existing Internet, in much the same way the original Internet layered itself on top of (that is, used the services of) the existing network at that timethe telephone network. At the same time, more recent considerations initially missing from early Internet designs,2 such as mobility (of nodes and of content location) and security (including content protection and trust), can be more holistically incorporated.
The following paper by Jacobson et al. presents a bold vision for such a content-centric network (CCN), presenting a design philosophy, architecture, and number of specific mechanisms. CCN requests for content indicate the name of content of interest (but not a location); CCN nodes are then collectively responsible for finding that named content and returning it to the requestor. One can discern lessons well-learned from the Internet in CCN's design, including the adoption of a best effort service modelallowing a content request/response to fail occasionally (perhaps due to node failure or mobility-related topology changes) and performing recovery at the end systemsand using hierarchical aggregatable names and longest prefix matching to scalably route requests to named content.
One can also discern the influence of more recent architectural innovations in the networking research community, including an emphasis on control by the receiver (who is, after all, the ultimate consumer of the content and the one therefore best positioned to make a decision) rather than sender control, and dynamic multipath transfers. And, of course, one sees innovations resulting from named content items being the unit of data transferin-network caching, breadcrumbs that plumb a reverse data path, inherent balancing of local node-to-node request/response flows, and more. We also see content security (including integrity, access control, trust, and provenance), network security (denial-of-service) and mobility all addressed in this seminal paper.
The architectures of large-scale networks (whether for communication, electrical, or transportation) seem to evolve in a style of punctuated equilibriumslow evolutionary changes between short periods of profound change. The telephone network evolved from human control to wired-logic control to stored-program control of electronic switches to today's IP-based control networks. Given the long lifetime of today's Internet architecture, the intense interest in future Internet architecture in the U.S., the E.U. and Asia (including various proposals for content- and information-centric networking), we may well be approaching a point of rapid change in communication network architecture. The following paper offers a perspective on the past as well as glimpse of what the future may hold.
©2012 ACM 0001-0782/12/0100 $10.00
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from firstname.lastname@example.org or fax (212) 869-0481.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2012 ACM, Inc.
No entries found