sshnet / sshnet/SSH.NET

Low throughput for sFTP over high latency high bandwith network

Open
#1,122 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C#
Stars
4.4k
Forks
993
Avg merge
9d 21h
Merged PRs (30d)
1

Description

Our client (that uses ssh.net) runs on AWS and the sftp server is hosted on Prem, and our uploads are multi-threaded.
We use the default UploadFile() method, which through various tests seem to perform the best on this network.

We however are seeing that the performance is way below par when compared to the sftp command from openssh for a single thread.
Like 3 times slower.
And with multiple threads, it goes crazy slow, to the point that the total throughput from all the threads are lower than that it gets from a single thread.

OpenSSH's sftp command on the other hand scales quite well.

Open SSH has a patch called HPN-SSH for such networks, not sure if something can be inspired from that.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the SFTP client’s default UploadFile() path and profile a single upload over a high-latency network, then compare it with the OpenSSH sftp behavior described in the issue. Reproduce the multi-threaded case and measure whether aggregate throughput decreases as threads are added; done means the cause is identified and throughput is improved or the limitation is documented.

Written by the indexing model from the issue text.

Assessment

Tech stack
csharp
Domain
networking, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.