Enabling a Faster Global Internet via Satellite - Industry Trend or Event

Telecommunications, Feb, 2001 by Ashok K. Thareja

Providing delivery solutions at both ends of the Internet is critical to tackling the World Wide Wait.

There is no lack of familiarity with the World Wide Wait. Even with the greater reach and lower costs of high-speed Internet connections, the number of Web sites--and the increasing volume of users accessing them--make it hard for networks to keep up.

Service providers of all types must find innovative ways to maintain their services and respond to the overwhelming demand from users.

Most Web content originates in just a few regions of the world (with the majority of it coming from North America and Europe) but is accessed by individuals around the globe. That's changing. A new report by Datamonitor predicts that by 2005, the origin of Web content will have steadily shifted toward markets outside the United States, creating a significant demand for high-speed and high-quality global content management and distribution solutions (see Figure 1).

Yet, the cost of bandwidth, infrastructure and the service itself depends heavily on location, and users in remote areas experience higher than normal fees and surcharges. Access to T1 lines in the most far-reaching areas can amount to tens of thousands of dollars per month--sometimes 100 times the amount it costs to provide the same service to more metropolitan areas.

Such constantly changing and uneven traffic patterns lead to slow download times and inconsistent delivery quality. As a result, service providers, particularly those outside North America and Europe, are forced to reconcile their distribution strategies to guarantee the most cost-effective and reliable means of service possible. Unfortunately, high bandwidth, infrastructure and service costs can hinder these intentions, leading to compromised QoS.

In the past year, however, new technologies have been developed to help alleviate the problems of congested terrestrial lines. While many of these solutions streamline Web pages for easier, landline transmission, another solution is to use satellite transmission capabilities to leapfrog Internet-based traffic (see Figure 2).

Look, Up in the Sky

The Datamonitor report predicts that demand for satellite-based solutions will soar by 2005. This isn't surprising: Satellite strategy requires minimal physical infrastructure, allowing the technology to be rolled out quickly to a broad area. Satellites also enable a stunning army of technologies that give service providers greater control and flexibility.

The most sophisticated satellite-based content distribution platforms combine intelligent protocols with a dual-edge strategy to address delivery problems for both content publishers and ISPs. Providing delivery solutions at both edges of the Internet is critical, otherwise it's a bit like driving halfway to work and walking the rest.

The basic idea behind intelligent control is to connect content providers in a few countries with end users in all countries. This is accomplished by establishing a primary cache at the content provider endpoint of the satellite link and a secondary cache at the end user side. This model involves multiple satellite links and primary and secondary caches. These two endpoint caches include a network of servers and routers that are scalable as the number and bandwidth of satellite links, Content providers and end users grow.

The primary cache in such an architecture is a large repository of Web content connected to the content provider edge of the Internet by a high-bandwidth link. The hardware uses this content publisher's link to gather interesting Web content in its cache. It also incorporates adaptive refresh techniques to make sure that the content is continually up-to-date. The content is continually scheduled for transmission via satellite and multicast to a group of secondary caches that subscribe to the primary caches.

The centerpiece of the secondary caches architecture, which resides at the service provider edge, is also a Web cache. Web content multicast by the primary caches is received, filtered for subscription and subsequently pushed into the secondary caches. All user requests for Web pages stored in the secondary caches are transparently intercepted and redirected to the cache, where the request is immediately served. If a request is received for content that is not stored, a terrestrial line is used one time to transmit the Web page to the secondary cache. After this initial request, the content remains stored in the secondary cache for immediate availability.

Three Key Strategies

Such a system relies on a combination of intelligent caching, multicast transmission and content management strategies to provide a fast, reliable path from one edge of the network to another. Intelligent controls reside in the primary and secondary caches and are essential to sharing limited satellite bandwidth across many users. A distributed algorithm (part runs on the primary caches, part runs on the secondary caches) facilitates communication between the hardware to enable it to communicate and coordinate resource use. Use of secondary caches minimizes the delay when content is requested, while the primary caches eliminate the need to burden the content providers' server with requests, thus saving bandwidth for content publishers.