Skip to content

Conversation

vkuznet
Copy link
Contributor

@vkuznet vkuznet commented Apr 15, 2021

This PR contains changes to DBSWriterModel to enable gzip payload reads. I adjusted the codebase to work with both types of payloads, the current one (plain) and gzipped payloads.

The gzipped payload can be supplied in HTTP request by using "Content-Encoding: gzip" header, e.g.

curl -H "Content-Encoding: gzip" -H "Content-type: application/json" --data-binary @/Users/vk/Downloads/bb.json.gz
http://localhost:8080

Once this code will be in place the DBS clients, like WMAgent, etc. can start supplying gzipped payloads to DBS POST APIs, like bulkblocks, blocks, files, datasets, etc.

The proposed changes are compatible with both 2.X and 3.X python versions, and fully support current mode of operations (without gzip) for payloads.

@vkuznet
Copy link
Contributor Author

vkuznet commented Apr 15, 2021

@yuyiguo , @amaltaro , @klannon , @KatyEllis I suggest that you take this PR into consideration as it can significantly improve our usage of DBS APIs and reduce latency related to large payloads. The changes are for DBS Writer, but once it is deployed, the other changes will be required to the clients, like WMAgent which should start adopting gzipped payloads and make adjustment to DBS POST API usage. I provided an example of how client should use HTTP request and it should be trivial to implement this in DMWM since it will only require to add extra HTTP header and start using gzip for payloads.

@yuyiguo
Copy link
Member

yuyiguo commented Apr 15, 2021

@vkuznet
This is on my to-do list. I'll let you know when I get into this.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants