Added article on multi site designs in metal-stack - #328
Draft
Sven-Ric wants to merge 5 commits into
Draft
Conversation
✅ Deploy Preview for metal-stack-io ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
mwindower
reviewed
Jul 30, 2026
mwindower
left a comment
mwindower
left a comment
Contributor
There was a problem hiding this comment.
Thanks for setting this up! I have first review comments and suggestions.
Comment on lines
+31
to
+37
| This is the default and most common pattern. Each zone hosts its own metal-stack partition, and partitions are fully independent. | ||
|
|
||
| No EVPN or VXLAN overlay is stretched between zones. Each partition operates a self-contained CLOS fabric with its own VNI space, BGP underlay, and exit switches. Private networks are allocated within exactly one partition, so a project that spans multiple zones receives a distinct private network per zone. All inter-zone communication is routed at Layer 3 across the exit switches, which exchange routes with external routing domains. | ||
|
|
||
| Partitions share no network state. Failure isolation is strict, and even the complete loss of one partition leaves the remaining zones unaffected. In return, cross-zone redundancy must be implemented at the application layer. Stateless services can be announced from multiple zones via anycast addresses, while stateful services require replication or state persistence outside a single zone. | ||
|
|
||
|  |
Contributor
There was a problem hiding this comment.
I would also mention here that each partition has its own storage system.
|
|
||
| The EVPN fabric is stretched across multiple data-center locations, forming a single partition. | ||
|
|
||
| A single VXLAN overlay spans all zones, which share one VRF and VNI space. Each zone hosts its own leaf switches, and the zones are interconnected at the spine or exit-switch level. The validated design covers three data centers whose interconnect forms a ring, which for three sites is equivalent to a full mesh. Interconnect topologies for more than three data centers are not defined. All data centers must lie within 20 km of each other, with latency below 1 ms, and all transport links must carry jumbo frames (MTU 9216) for VXLAN encapsulation. |
Contributor
There was a problem hiding this comment.
The storage system is here distributed over several locations.
|
|
||
| ### Design 3 - Zone Aware Setup (Roadmap) | ||
|
|
||
| This design targets a middle ground between independent partitions and the metro setup. Multiple independent partitions exist in different locations and EVPN is not stretched between them. Instead, the exit switches interconnect selected VNIs across partition boundaries, for example via EVPN Multi-Site DCI or SRv6. A private network in one partition thereby becomes reachable from another partition without traversing external networks. |
Contributor
There was a problem hiding this comment.
I would add Route-Leaking with a DCI-VRF as third implementation example there.
…opologies.md Co-authored-by: mwindower <markus.fensterer@gmail.com>
…opologies.md Co-authored-by: mwindower <markus.fensterer@gmail.com>
…opologies.md Co-authored-by: mwindower <markus.fensterer@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
WIP article on multi site architectures in metal-stack.
TODO
Used AI-Tools ✨
Research supported by fable 5