Features/Asynchronous NBD: Difference between revisions
(→Status) |
No edit summary |
||
| Line 14: | Line 14: | ||
== Status == | == Status == | ||
Committed. | |||
== Links == | == Links == | ||
* [http://github.com/bonzini/qemu/commits/nbd-trim git tree] (branch nbd-trim, for the client) | * [http://github.com/bonzini/qemu/commits/nbd-trim git tree] (branch nbd-trim, for the client) | ||
* [http://github.com/bonzini/qemu/commits/split-main-loop git tree] (branch split-main-loop, for the server) | * [http://github.com/bonzini/qemu/commits/split-main-loop git tree] (branch split-main-loop, for the server) | ||
Revision as of 14:57, 13 January 2012
Summary
Supporting asynchronous I/O in the NBD block device and qemu-nbd server makes it easier to use high-performance remote storage.
Owner
- Name: Paolo Bonzini
- Email: pbonzini@redhat.com
Description
The NBD block device and qemu-nbd server provide a simple base for using remote storage within QEMU. NBD can also be coupled with other features such as streaming or copy-on-read. However, the NBD implementation in QEMU is entirely synchronous, thus limiting guest performance and being hardly applicable when multiple guests use the same NBD image (for example as backing storage for QCOW or QED).
Status
Committed.