I'd like to boot different operating systems over the network but the PXE client should not be directly connected to the system hosting the filesystem mostly for security reasons. My idea was to use some kind of middle man/proxy which sits between the client and the host. The middle man for the filesystems, however, is really limited in hard disk space (couple of 100 MBs) and it is currently not possible to increase that. Most filesytems are way above this size. I was thinking of adding only small chunks of the filesystem to the hdd of the middle-man e.g. via network, manually copying, rfc1149 etc. The pxe client then loads one chunk, chunk gets removed on the middle man and the next chunk is added, client loads second chunk etc. In the end the chunks are joined on the client and the complete filesystem is loaded. Does anyone here know if this is possible?


For clarification the final system could look like this:

Filesystem-host/Super-Server (splits filesystem) --> middle-man (receives one chunk at a time from the host and serves it to the client) <--> PXE Client (joins chunks and boots the filesystem)

To join the chunks and boot the complete filesystem is the hardest part. I could not find anything on the net related to that so I'm not sure if this is possible.

  • I’m not sure I understand what you mean by “the host for the filesystems”. Where would the filesystems actually be — some sort of super-server?  Wouldn’t it be simpler to have the clients talk directly to the super-server? Sep 13, 2017 at 11:22
  • Yes super-server would fit quite well in this case and of course it would also be easier for the client to directly talk to this server. However there are reasons (mostly security releated) that there is this middle man and the client does not talk directly to the super server. If this split and join stuff is not possible I have to come up with something else.
    – whyisthat
    Sep 13, 2017 at 11:35
  • 1
    Please edit your question to include the essential information from these comments. Sep 13, 2017 at 13:05
  • I'm confused about how exactly the middle-man is supposed to increase security, and why it would need a harddisk to do that. Can't the middle-man just be a firewall that blocks direct access, but let's the PXE packets (or at least the TFTP packets) through? In the end, no matter what you do, the middle-man reads from the server, and sends data to the PXE client. (It's also not particular difficult to set up the TFTP server on the real server very securely, without any middle-man).
    – dirkt
    Sep 13, 2017 at 15:01
  • The setup is certainly not the most convenient one and also might make only sense to me :) but at the moment there are no other easy options. Still the main question about how to pxe boot a split filesystem remains unanswered.
    – whyisthat
    Sep 13, 2017 at 20:14

1 Answer 1


your approach is not good. Can you imagine your "middle-man" dealing with let say "5" simultaneous clients installing multi GB OSs?

A middle-man is not the way to add security to PXE. If you are still in doubt just consider that no-one in the market does such a thing.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .