I took a look at the StarlingX Overview Slide Deck.
Some ideas on updates that could be done:
-
Add Geo-Redundancy of Central Cloud in Distributed Cloud configurations
-
SLIDE 6
-
@Ngo, Tee do you have a
Geo-Redundancy diagram that you could share with Ildiko ?
-
SLIDE 9 - StarlingX - Edge Virtualization Platform
-
i.e. very very high-level software architecture diagram
-
changes ?
-
under OpenStack Components
-
remove ‘Bare Metal’ and ‘Object Storage’ blocks
-
under Kubernetes Components
-
I would change ‘DNS Service’ to ‘Core DNS’ … I believe that is what it is referring to,@Webster,
Steven ? and
[> steve]: I think that’s fine ^
-
What is Network Proxy referring to ? CNI ?
[> steve]: Yeah this is kind of a weird one. I think it would be much better referred to as you say below by “CNI”.
Should we just change that to ‘CNI’ or ‘Container Networking’ ?
@Webster, Steven ?
[> steve]: I think “CNI” is fine to keep as it’s commonly used in the community and it probably makes it easier to fit on the diagram.
-
We should similarly add ‘CSI’ or ‘Container Storage’
-
We should change ‘Helm’ to ‘FluxCD, Helm’
-
We should change ‘Dashboard’ to ‘Kubectl , Dashboard’
-
I would add a block for ‘Integrated OpenSource Container Services’
-
i.e. for all the open-source containerized services that we integrate into solution
-
e.g. nginx, cert-manager, metrics server, istio, vault, dex, kubevirt, security profiles operator, portieris, etc.
-
for the layer just above Linux
-
change to ‘Linux - Debian’ ?
-
Change the ‘Kubernetes + OpenStack’ block to just ‘Kubernetes’
-
OpenStack is in the Containerized block above it
-
Remove ‘TPM’ block … not supported anymore
-
we have plans to add it back in … but not sure when that will happen
-
Change ‘Low Latency Linux’
-
to: ‘Standard or Low Latency Linux - Debian’
-
For scalability bullet … should replace with something like
-
Scalable deployment models
-
Support 1-200 k8s workers
-
Support 1-10 k8s workers configured as OpenStack computes
( maybe don’t say this one … its embarrassing 😉 )
-
Support 1-1,000 Subclouds
-
where each Subcloud is a cloud of 1-200 k8s workers
-
The bullet on “Focus on minimizing the infrastructure footprint” is not well written … not understandable
Suggested REWORD
-
Minimize infrastructure footprint / overhead
-
For Standard (Controllers + Workers) Configuration
-
support 2-server controller/master solution
-
on worker nodes, only 1 core required for StarlingX Platform
-
For All-In-One (AIO) Configurations (i.e. AIO-Simplex and AIO-Duplex)
-
AIO servers require only 2 cores for StarlingX Platform
-
SLIDE 20 “Kubernetes Cluster Software Components”
-
Change Armada to FluxCD
-
Personally I would change ‘Ingress LB’ to ‘Nginx’ … to avoid any confusion
-
I would remove the current and future stuff
-
SLIDE 26 “Application Management”
-
Need to change the Armada stuff to FluxCD
-
@Reed, Joshua ? Could
you provide updated text and diagram for this slide ?
-
SLIDE 29 Supported OpenStack Servies
-
Ironic and Telemetry(Ceilometer Gnocchi Panko Aodh)
-
… not currently supported.
-
SLIDE 32 & 33 … on Distributed Cloud
-
@Ngo, Tee ? can you provide
updated text for these two slides … and diagram if you feel its required
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.
I’m reaching out to you about the onboarding/overview slide deck[1] that we have on the Learn tab of the project’s website.
The slide deck has a lot of diagrams, including ones that describe the Flock services, ones that go into the Kubernetes-related parts of the platform, as well as a couple of high level overview diagrams. It came up in the past that most
of these diagrams need to be updated, since the platform has evolved quite a lot since they were originally drawn.
I can update the diagrams in the slide deck, and work with the Foundation’s design team to get the ones updated that we use on the website and in the one-pager. Does anyone have any newer diagrams that you could share for guidance? Or,
should we utilize some time on the community call to talk through which diagrams need updating and find volunteers who can sketch updated versions of them?
———
Ildikó Váncsa
Director of Community
Open Infrastructure Foundation