Can Riverbed save the internet from Doomsday?
AT&T recently warned that the internet will run out of bandwidth in two years. It sounds crazy, but Jim Cicconi, AT&T's VP of Legislative affairs says that the current systems that constitute the internet will not be able to cope with increasing amounts of user generated content and video. He went on to say that in three years time, 20 typical households will generate more traffic than the entire internet today!
When you think about it, this doesn't sound so far-fetched.
Video on demand and video streaming is becoming more ubiquitous with the success of sites like YouTube, Joost, and Tvunetworks. Not only are there more people using video on demand technologies, but the content quality is also improving, file sizes are growing at a staggering rate, and the bandwidth requirements are going through the roof. Are we seeing the perfect storm for massive plumbing problems taking place across the biggest set of pipes that constitute the internet as we know it today?
So what would it take to address this looming issue? AT&T says it will require an infrastructure overhaul costing over $55 billion over the next 3 years in the US alone!
I have a better solution. Riverbed's Steelhead appliance and Steelhead Mobile software. I know it sounds wacky, but Riverbed can save the internet! Hear me out for a minute...
Riverbed's Steelhead appliances can be installed at every point of presence where videos and user-generated content is sourced from. Steelhead Mobile software can then be installed on consumer's desktops or laptop PCs where the content is requested. For corporate traffic, Steelhead appliances can be installed at the data center and branch office with Steelhead Mobile software installed on mobile workers laptop or desktop machine if they work at home.
Now that we have a Riverbed Steelhead Internet overlay, how is this going to save the internet? The idea is simple. The RiOS software that powers the Steelhead appliances and Steelhead Mobile software has a powerful data deduplication feature called data streamlining, which effectively eliminates all redundant TCP traffic from the network. For video on demand traffic, which is estimated to be 80% of the internet's traffic in the near future, eliminating the redundancy could be huge.
Let's look at the YouTube example. A home user with Steelhead Mobile on their PC clicks on the YouTube browser to view a popular YouTube video. The Steelhead Mobile software will automatically discover and peer with the Steelhead appliance that is configured where the YouTube video is sourced. If that video was accessed by that user sometime in the past then the Steelhead appliance on the YouTube source side will tell the Steelhead Mobile software that it has seen the data that constitutes the video before and to reconstruct that data from its local datastore and deliver the video locally instead getting all the data over the internet. That 10MB video was accessed over the internet, but a pointer to the video allowed the video to be reconstructed locally. In this example a tiny reference consisting of a small number of bytes represented 10MB of data. Only the tiny reference had to go over the internet!
Think of the bandwidth savings. This groundbreaking deduplication technology is applied to all TCP traffic so all the email, file sharing, and web applications benefit as well. Unlike video, this other traffic can also be reduced in size 50% or more during the first pass of seeing the data and between 60 to 95% after seeing the data previously . This combined with eliminating redundant video traffic results in the potential for a huge bandwidth savings.
Can we get a trial going? Al Gore may have invented the Internet, but Riverbed can save it.
( http://news.cnet.com/2100-1034_3-6237715.html ) and various articles covered in the past few months have all been focused around how Ineternet usage and bandwidth would be 50 times more than today and the content created then would jam the internet space.
Typically utilization of bandwidth and usage is directly proportional to the no.,quality & results of QUERIES that a user inputs from the moment he logs on. These queries can be as simple as searching, or as intense as uploading. At the end of it, its all about minimizing the no. of queries.
We are trying to research the potential ways of minimizing these queries by pre-defining most searched “keywords” & “categories”.
By not entering any search query, or typing for searches you would be skipping multiple steps and help in saving energy and bandwidth.
If we all try and add to this pre-defined list of searches, we can help save far more energy, bandwidth and money.
Help us all move towards the green technology.
http://www.lazii.com
(currently in beta stage)
Posted by: Ankur Dewani | September 15, 2008 at 03:18 AM