A distributed system consists of multiple nodes that share incoming load among themselves. In a centrally coordinated environment, the responsibility falls on a gateway to manage request dispatch to this cluster of nodes. In this paper, we propose an adaptive and unified approach to request dispatch to enhance decision making capabilities of such a gateway in case of partial and total cluster shutdown states. Using the described algorithms as a base, we derive a formula for landing probability on a node and explore its implication on various cluster states. We also show, via graphs, the request dispatch pattern when using this approach with adverse cluster states.
-
Notifications
You must be signed in to change notification settings - Fork 0
Accompanying white paper for idea behind ServiceQ (https://github.com/gptankit/serviceq)
gptankit/serviceq-paper
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
Accompanying white paper for idea behind ServiceQ (https://github.com/gptankit/serviceq)