>> continued from previous page
Putting it Together
Solution providers incorporate eCDNs in package deals. Thes packages comprise three core technologies: cache servers, content distribution and content redirection.
• Cache Servers. Cache servers reside close to original content servers or near end users on remote WAN links (see our online-only sidebsar, "Proxy Primer"). These devices dynamically pull content from origin servers based on end-user requests, then maintain that content in cache for later use. They also can store pre-positioned content copied or mirrored from central servers. Their sole purpose is to speed content to end users. Since they act on behalf of origin servers, they are often called proxy caches.
Proxy caches do far more than store and serve static Web pages. They support a variety of file types and content from FTP and NNTP (Network News Transfer Protocol) servers, as well as streaming media from Apple QuickTime, RealNetworks RTSP (Real-Time Streaming Protocol) and Microsoft Metadirectory Services. They can serve streaming media on demand or live. For live events, caches support stream splitting, in which a single streaming file is received over a WAN link and split over the LAN to downstream users. This one-to-many delivery mechanism leverages multicasting in the enterprise (see "The Wizardry of Multicast,"). Caches also include built-in functions or algorithms to ensure fresh content based on TTL (time to live) and other factors, all of which can be set using content distribution technology.
• Content Distribution. ECDN technology ensures that cached files are current and synchronized with the files on the origin servers. A central controller distributes content from server to cache by pre-positioning the files prior to end-user requests. Pre-positioning can be achieved by mirroring or copying content through an administrative action or by pushing to or pulling from origin servers on a schedule.
Once the controllers distribute the content, they can use policies to manage it. Basic policies can "pin" popular content in cache and make sure it is always served from memory. They can also dictate refresh rates by TTL factors or by date and time. Controllers can refresh or purge caches on demand or according to a schedule. These policy controls also allow local optimization and personalization for the enterprise.
Distribution policies can determine which content resides at which enterprise locations. If your users read newspapers every morning, you can pre-position those papers in cache, where they can be read on demand. If you use streaming media for corporate communications, you can position the files on multiple caches and let users view them according to a schedule you set. Pushing popular and premium content to certain locations reduces latency and network bandwidth constraints at peak hours.
Policies even help secure cached content. ECDNs can integrate with directory and authentication services schemes to provide access control to cache servers and content. For example, Volera's Excelerator cache server supports LDAP, NDS, RADIUS and certificate-based authentication, while CacheFlow's cIQ Content Delivery Architecture supports LDAP, NTLM (NT Domains) and RADIUS.
Content-distribution controls do not compete with content-creation and -management tools, such as those offered by Interwoven and Documentum; however, vendors are workin gon interoperability. For example, InfoLibria's technology is tightly integrated with Microsoft Producer and lets content creators publish directly to cache servers.
• Content Redirection. The third element of eCDNs, content redirection technology, lets content move to the closest available caching server to fulfill end users' requests. ECDNs use local or global redirection.
Local redirection techniques include explicit proxy directives in applications and intelligent switching or routing devices, and work in transparent or nontransparent mode. In nontransparent mode, an end user's browser or media player must be configured to point to the cache. Local redirection in transparent mode uses intermediate devices with caching servers. WCCP (Web Cache Control Protocol)-enabled routers or Layer 4 to Layer 7 switches redirect requests for original content to caches. Using intermediate devices in transparent mode works well if every branch office has its own caching device. Otherwise, global redirection is required.
Global redirection directs end-user requests for original content from a central server or global redirector. DNS servers can act as global redirectors at the domain level. But DNS redirection does not distinguish requests by content type, location or resource availability. The redirected resource may be taxed already, so vendors provide smarter solutions to route requests. For example, InfoLibria's Content Director includes global redirection functions that direct requests to specific caches based on content, service location and resource availability.
If you have no need to stream media from central servers, check out the cache servers included with each of the eCDNs we tested; caching frequently used Web objects can cut costs. Then, if you begin delivering streaming media files, you can investigate the other pieces in the eCDN space and send your remote offices and branch locations into warp speed.
Sean Doherty is a technology editor and lawyer based at our Syracuse University Real-World Labs®. A former project manager and IT engineer at Syracuse University, he helped develop centrally supported applications and storage systems. Send your comments on this article to him at sdoherty@nwc.com.
RSS
