[tahoe-lafs-trac-stream] [Tahoe-LAFS] #3939: HTTP protocol is significantly slower than Foolscap protocol

[Tahoe-LAFS]

#3939: HTTP protocol is significantly slower than Foolscap protocol
--------------------------+-----------------------------------
     Reporter:  itamarst  |      Owner:
         Type:  defect    |     Status:  new
     Priority:  normal    |  Milestone:  HTTP Storage Protocol
    Component:  unknown   |    Version:  n/a
   Resolution:            |   Keywords:
Launchpad Bug:            |
--------------------------+-----------------------------------

Comment (by itamarst):

 Looking at an immutable, here are some writes for a (presumably small)
 object, as recorded via layout.py: 36 bytes, 1400, 32, 32, 32, 170, 320.
 This will get batched via pipeline.py (see #3787) and for Foolscap maybe
 that's fine. But HTTP/1.1 has higher overhead per query than Foolscap I
 suspect, and even with HTTP 2.0 I imagine it's rather higher.

 So possibly one strategy is doing an alternative to `pipeline.py` where
 logic isn't generic "batch API queries" but instead it relies on the fact
 we're doing writes without any holes, so we can coalesce writes
 semantically. (Previously it _would_ have holes so this would've been
 harder.)

--
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/3939#comment:2>
Tahoe-LAFS <https://Tahoe-LAFS.org>
secure decentralized storage