Web Application Acceleration

On the home page users are frequently presented with latest media list, along with “most popular” section and there may be some sponsored content. In addition to regular search screen, these site often provide folding menus with “other media like this”, “other media from the same author”, “people that saw this media file also were interested in these other media files” and so on.

Something common to media sites is spiky, uneven user traffic subject to massive swings. Media assets, such as videos have a tendency to explode in user popularity in a few short minutes, as the buzz spreads around the Internet. Popular news aggregators pick up the new content and users start notifying each other about the newest and the hottest video.

There’re are many examples of viral, incredibly popular videos that caught the industry by complete surprise. No media expert could have possibly predicted just how popular a given video will become.

No matter the content, popular videos are guaranteed to test the performance limits of even the most powerful web sites. This is where aiCache ability to deliver, instantaneously and with zero overhead, web responses to the visitors can really save the day.

Starting from the home page, to search results, to media metadata and other API results, content can be easily cached, removing truly enormous amounts of traffic from your origin infrastructure. Just a pair of aiCache servers can deliver in excess of 200 Million page views per hour, while trickling just a small percentage of the traffic back to your web, application and database servers. Enough capacity and throughput to deliver even in the most explosive of web traffic situations.

Another common element that is frequent on modern media sites is user generated content. Most noticeably the comments and feedback that viewers leave for media assets. This present a challenge of its own – how do you configure the site so that the user comments are always fresh without placing a huge burden on site’s persistent stores. The persistent store is normally a Relational Database, frequently front-ended with Memcached DB results caching solution.

The solution is to intercept and respond to vast majority of user requests before they ever reach your web and persistent store component. How do you guarantee that comments stay fresh and the newest comments are displayed right away. How can be sure that the poster is not confused and sees their post right away so they don’t repost the comments over and over again ?

aiCache offers a unique feature aimed at solving this very problem: response driven cached content invalidation. All you need to do is to modify the code that receives user comments to return, in its response header, an aiCache expiration command. Voila, problem solved.

Now, as long as there’re no additions to the content, it is always served directly from aiCache, never burdening the site’s infrastructure. When new content is posted, the cached response is instantly refreshed, reflecting the newest submission. When deployed in clusters, aiCache serves communicate to making sure that content expiration and refresh happen synchronously.

To see for yourself how aiCache can benefit your site, please call us to discuss your site’s needs. You can also download an obligation-free, fully functional version of aiCache for a quick, easy-to-use and non-disruptive trial.