View Single Post
Old 09-03-2008, 02:03   #832
Phormic Acid
Inactive
 
Join Date: Mar 2008
Services: Still to decide on Aquiss or Be
Posts: 62
Phormic Acid is an unknown quantity at this point
Re: Virgin Media Phorm Webwise Adverts [Updated: See Post No. 1, 77, 102 & 797]

Quote:
Originally Posted by bigbadcol View Post
If it is deleted after a few hours then it must be stored. Somthing that has been denied most strongly over the past few days.
I also think Phorm’s claim that no information gets written to disc is probably rather misleading. When people look at their computer at home and think “where am I going to store 500GB of data?”, the answer will be “on a hard disc.” If you’re a business, this isn’t necessarily so. You could get something like the Gear6 CACHEfx.

Gear6 satiates hungry apps with 500GB RAM monster

I’m not suggesting Phorm will be using that. They’ve probably built their own storage system using standard parts.

Phorm have only got to store textual pages. So, not memory-hogging images and videos. Let’s assume the average HTML page is 100KB in size. You could get around 5 million pages into 500GB, allowing for storage overheads.

Many big pages contain a lot of junk. They’re generated on request, using things like PHP. They’ll contain the same code snippets over and over again, lots of white space or even large numbers of HTML comments. That could all be stripped out prior to caching. It could be as simple as stopping passing on the stream when a ‘<’ is reached and starting again on the next ‘>’. You might want to collect the alternative text for images. That would add a bit of complexity, but there’s no reason why it can’t be done in real-time within the stream, saying using some FPGAs.

If you strip out all the crud, you might get down to an average of 25KB per page. Then, you’d get around 20 million pages into 500GB of RAM.
Phormic Acid is offline