NetOp Systems Fead Optimizer: Mind-blowing Compression

Netopsystems FEAD Optimizer uses a high compression technology (FEAD). What's so special about this technique is that it can compress already compressed files by 20%-40%. Adobe Reader was one of the first to incorporate FEAD Optimizer. At first look the decompression termed as 'recomposition' by Netopsystems, looks intriguing, being heavy on the CPU and taking quite a lot of time and the absence of cancel button to stop the decompression (Oops! Recomposition) is further displeasing. But, the compression ratio of FEAD is more than enough to overcome these drawbacks. Here's a comparison between FEAD and a few well known compression algorithms.

Adobe Acrobat Reader 7.0

Total Files---> 8
Total Size---> 32.7 MB
Portable Zip---> 30.2 MB
Enhanced Deflate---> 30.2 MB
Bzip 2---> 30.3 MB
PPMD---> 30.7 MB
LZX---> 30.1 MB
FEAD---> 19.8 MB

Panda Antivirus Titanium 2006
Total Files---> 321
Total Size---> 80.4 MB
Portable Zip---> 39.4 MB
Enhanced Deflate---> 38.9 MB
Bzip 2---> 38.0 MB
PPMD---> 35.0 MB
LZX---> 32.0 MB
FEAD---> 22.2 MB

As you can see, the FEAD technology is much more effective in compressing pre-compressed files. It is even better than the best these days (PPMD and LZX) by some 30% (Approx). UPX is the only algorithm I have heard about that can compress Portable Executables by that much. Then I learned that it was indeed a modified version of the UPX algorithm which they have named NOS. Some people reckon, that in doing so they have violated the GPL. Check this out for more details.

Possible technique: Netopsystems have come out with a basic but effective technique to copy repeated files just once in the source package and copy them as many times required during recomposition in the destination folder(s). There must be many other things they have put into use, because UPX can only compress PEs which in the case of Panda Antivirus is 10.5 MB in size. Had they compressed the 10.5 MB PEs to 0 Byte, that would have accounted for the 10 MBs deficit (which even by Napolean's standard is impossible) . So, they must be using some other algorithm for other files.

Another probability: They haven't released their FEAD Optimizer for users, rather they rely on customers handing their Setup files to them for analysis and compression. It might be this step that hands them the edge. Maybe they search the files for every drop of fatty byte they can shed.

Whatever be it, the company has opened doors for further research in the field of compression algorithms, which was seeming to have reached its zenith.

 

 

 

Comment/Reply (w/o sign-up)

Panda Internet Security has done it again. They have continued with the FEAD Optimizer they employed in the 2006 edition of their product. The FEAD optimizer seems to have been upgraded as the new version finally has the cancel button which was painfully lacking on the previous one. Also comes the handy ability to pause the recomposition process.

As before I'll offer the same compression statistic:-

Panda Antivirus Titanium 2006
Total Files---> 289
Total Size---> 90.6 MB
Portable Zip---> 45.2 MB
Enhanced Deflate---> 44.6 MB
Bzip 2---> 43.9 MB
PPMD---> 40.7 MB
RAR (Good) ---> 38.9 MB
RAR (Best) ---> 38.9 MB
LZX---> 38.0 MB
RAR (Good + Solid) ---> 36.1 MB
RAR (Best + Solid) ---> 36.0 MB
FEAD---> 28.7 MB

Comment/Reply (w/o sign-up)

what about rar and 7zip ?

Comment/Reply (w/o sign-up)

So this compression method is only available to the software vendors and not to the public? I have used UPX before and love it because it does have a big compression ratio for my program. If FEAD is available for public use, I would love to give it a test run.

The link you posted is a dead link. Google has a cached page of it though. Just do a search for Netopsystems FEAD Optimizer in Google and a few links down, you will see "GPL violation in Netopsystems FEAD Optimizer/Recomposer used by ...". Instead of clicking on that link directly, click on the Cached link underneath it to read it.

Comment/Reply (w/o sign-up)

At the time I made those earlier statistics, I didn't try the RAR compression format. Check out the latest statistic on Panda Anti Virus 2007. I have covered most of the formats there. 7zip is basically a format to store the compressed files in, it doesn't have a compression algorithm of its own.

Yes. I presume they figured that having a trial version out would've left them at the mercy of the crackers. I don't believe they have a completely automated process to package the setup. I have heard that they take the setup files from the vendors, study the files, apply appropriate compression to the files (like UPX for EXEs & LZA for text files), reduce size overhead due to duplication by having only single compressed code for such files.



Yeah, I have read that article on how they might have just re-named the header of the UPX format.

 

 

 

Comment/Reply (w/o sign-up)

But it still does a better job at compression data. If you check the 7-zip website you can see it compresses better as winrar (for the two examples).
http://www.7-zip.org/

Comment/Reply (w/o sign-up)

You are correct. I didn't imagine that 7-Zip could achieve better compression ratios than Win Zip, despite using the same compression algorithms - BZip 2 & PPMD. It also beats the RAR format with the LZMA format.

Here goes the statistics:-

7-Zip BZip 2 --> 41.9
7-Zip PPMD --> 39.5
7-Zip LZMA --> 37.2

7-Zip Solid BZip 2 --> 41.0
7-Zip Solid PPMD --> 37.7
7-Zip Solid LZMA --> 33.2

Comment/Reply (w/o sign-up)

FEAD is fine if you're really crunched for bandwidth, but the problem remains that you have a time/compression tradeoff. I remember trying to install Acrobat 6 on all the machines on my network at work when it came out...there were still 400 MHz machines up there that spent an ungodly amount of time unpacking that installer.

On the plus side, Adobe's antics (like using FEAD) are what made me seek out Foxit Reader for my PDF reading needs. There's even a U3 version.

Comment/Reply (w/o sign-up)