Question

I am writing a utility that will zip a file (or set of files) using the LZMA sdk then send the file off to a ftp server. Usually the speed of the compression is faster than the speed of the ftp connection. What I would like to do is instead of compressing the file, waiting for it to finish, then starting the upload I would like to compress to a temporary file or stream, then while it is being compressed upload the completed portions.

The question now is how?

One concern I have is the files I will be working with can be over 1GB when compressed and the systems I will be running this on will have between 512MB and 2GB of ram so I do not want to let the compression side run wild in to memory and lock up the system. The method I have been thinking about is running the compression in a thread, Queuing up 5-10Mb in a memory stream, then send the info to the ftp in the other thread. Is this a good approach or is there a better way to do it? Are there any gotchas like it needs to rewrite the file header at the start of the file when it is done or anything else?

I plan on writing this in c# but code examples in c, c++, or java are fine too.

Thank you for your help.

Answer 1

I’m not familiar with the LZMA SDK but in C# with the SharpZipLib library it’s easy to stream out zip files. You don’t have to worry about memory, only the blocks being compressed/streamed will be in memory at any one time. We use this to compress and stream files via HTTP but the concept for FTP is the same.

Basically you create a ZipOutputStream that passes data off to the FTP stream. Call PutNextEntry at the start of each file and then stream the file contents. Not much more to it than that.

http://www.icsharpcode.net/OpenSource/SharpZipLib/