Content Delivery Network

Social networks are integrated into our lives nowadays. We have a variety to choose from. While Facebook and MySpace define a category, Flickr, Tumblr form another and then we have Twitter, Friendfeed and likewise.

Evey time we have a friends blast or a reunion, we are happily posting snaps and videos over to them. But have you ever wondered where these files go?

Here comes CDN or Content Delivery Network. Most people have the idea that facebook, myspace, they all store the data into their own data servers. That is the wrong idea.

Networking giants like Facebook, MySpace, even Last.fm use Content Delivery Networks to store the files we upload. They tie up with companies dealing in the system and only keep a link to the actual file on their servers. So, when we are accessing a media file on someone’s profile, it is actually being delivered from a totally different server from that of the website’s server. A CDN thus, can be considered a P2P based network, only here, we need a browser to access data.

A CDN identifies two types of data request.

1. A Download request: Eg. A file download
2. A stream request: Eg. Seeking through a video

As the number of occurrences of the file in a network increases, the concurrency in the network increases. Thus we can expect better performance and higher availability of data. The best thing about CDN is it’s scalability. As the size of data increases each day, the network can adapt itself to accommodate more data in it. Also, CDN allows a level of bandwidth optimization as the actual data is simply pointed to.

But it also has it’s downsides. Here are three disadvantages of CDNs.

• Data in a CDN is open to itself. There is not much access control implemented on a CDN. Thus, if any one of the participants of a CDN’s becomes a public HUB, there is a fair risk of a huge database, possibly proprietary, being available openly.
• Data in a CDN is not updated as frequently as they are on the parent website. This is where claims of privacy comes in relating to Facebook retaining user accounts who have left, and many more. Actually it is not Facebook’s fault. Facebook has cared to afford enough to write data to the network, but has to pay an extra cost to remove data which is just not worth the cause. So it waits for an overwrite which is usually due in a couple of months.
• If you are participating in  a child organization, the parent organization can make your user data and statistics public without prior notice, like what happened with last.fm.