Skip to content

Commit

Permalink
Copy piece from the original doc.
Browse files Browse the repository at this point in the history
  • Loading branch information
Marcin Owsiany committed Oct 16, 2017
1 parent 07e214b commit 4553805
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion sig-scalability/goals.md
Original file line number Diff line number Diff line change
Expand Up @@ -77,6 +77,6 @@ Configuration of the control plane for cluster testing varies by provider, and t
## Open Questions

1. **What, if any, reasonable use cases exist for very large numbers of very small nodes (e.g. for isolation reasons - multitenant)? Based on comments so far, it seems that the answer is yes, and needs to be addressed.**<br>
The above scaling goals explicitly accommodate a maximum of 5,000 nodes. Do we need a special case for larger numbers of small nodes (e.g. 200,000 single-core nodes). The latter already fits within our other scalability limits (cores per cluster, pods per cluster), so it might not be more difficult to achieve than those. Two example use cases I've heard anecdotally are (a) running e.g. large numbers of customers' small, largely idle wordpress instances, one per node, and (b) giving away limited numbers of free
The above scaling goals explicitly accommodate a maximum of 5,000 nodes. Do we need a special case for larger numbers of small nodes (e.g. 200,000 single-core nodes). The latter already fits within our other scalability limits (cores per cluster, pods per cluster), so it might not be more difficult to achieve than those. Two example use cases I've heard anecdotally are (a) running e.g. large numbers of customers' small, largely idle wordpress instances, one per node, and (b) giving away limited numbers of free containers to large numbers of customers for promotional reasons (similar to AWS free tier micro instances).
2. **What, if any, reasonable use cases exist for very large numbers of very small containers per core?**<br>
E.g. are perhaps hundreds of containers per core useful for specialized applications? We speculate that building large numbers of very small yet useful containers each within say 20MB of RAM, and <1% of a core is difficult (as opposed to very small supportive/sidecar pods alongside larger pods, which is totally legitimate, and supported).

0 comments on commit 4553805

Please sign in to comment.