Tim Burke candidacy for Swift PTL (Caracal)
Change-Id: Ibd9ab22e4c9ea8d2d8a5695d73da85d5eafa11d4
This commit is contained in:
parent
0c17dc92c2
commit
987e9058fd
22
candidates/2024.1/Swift/tburke@nvidia.com
Normal file
22
candidates/2024.1/Swift/tburke@nvidia.com
Normal file
@ -0,0 +1,22 @@
|
|||||||
|
It would be my pleasure to continue serving as Swift PTL for the
|
||||||
|
Caracal cycle.
|
||||||
|
|
||||||
|
I forsee us grappling with several challenges over the next few cycles:
|
||||||
|
|
||||||
|
Successful clusters expand until they hit a wall, whether that's due
|
||||||
|
to budget or physical space constraints. Swift will need to adapt to
|
||||||
|
operate in more constrained environments.
|
||||||
|
|
||||||
|
Total disk capacity is not the only measure of a cluster. Network
|
||||||
|
bandwidth (both aggregate across the cluster as well as single-stream)
|
||||||
|
and requests per second drive much of the client experience. Swift will
|
||||||
|
need to be able to ration those between tenants, similar to how it can
|
||||||
|
enforce capacity quotas.
|
||||||
|
|
||||||
|
Operators are not the only ones interested in metrics and storage
|
||||||
|
insights. Users want to know what their individual access patterns
|
||||||
|
are like and how that has changed over time.
|
||||||
|
|
||||||
|
I'm excited to see how much we can achieve in Caracal.
|
||||||
|
|
||||||
|
Tim Burke
|
Loading…
x
Reference in New Issue
Block a user