Compression Classes Enhance I/O in .NET 2.0

CodeGuru: Compression Classes Enhance I/O in .NET 2.0

"... Specifically, it demonstrates reading and writing compressed data with the new System.IO.Compression namespace.

The System.IO.Compression namespace introduces a series of classes that provide compression and decompression capabilities. The classes exhibit similar behaviors and characteristics, but the compression classes operate against streams. To my knowledge, the .NET Framework 2.0 will include two algorithms for compression/decompression: Deflate and GZip.

..."

Another short posting about the new System.IO.Compression namespace.


In my mind, I'm still balancing the space saving of file compression verses convenience/performance.

I deal with millions of files or file fragments (email bodies in RTF/Plain text, etc), so saving a little file/db space can make a huge difference. BUT it is very nice to be able to see these files/fragments in their native form with first uncompressing them (for QC/troubleshooting/etc).

Right now space is not a major issue. Performance (in-out throughput actually) is king. Network performance is also not a major issue, so compressed bits on the wire is not all that attractive.

Also there are actions that require the uncompressed file/fragment (full text indexing, exporting, etc).

Now if I could get my primary consumer to accept compressed files (but then again, that system would have to decompress them for indexing, delivery, etc) so it's not likely.

Looks like I've talked myself out of it (for now). :/



Past Post XRef:
System.IO.Compression (.Net 2.0)
zVarchar example [SQL Server 2K5 .Net UDT]