Hi Marcus,

Thanks for the reply. I will follow up on your first two points:


  1.  I did not complain about performance or optimization. I complained about 
redundancy and reinventing wheels. What bothers me is that ONAP eats 100GB of 
space and what actually offers in return? Where is the functionality?
  2.  Maybe there are use-cases which ONAP can handle but are they usable in 
practice or are they just a one-shot demo? There is this article 
https://www.sdxcentral.com/articles/news/bell-canada-first-to-deploy-open-source-onap-in-production/2017/12/
 claiming that ONAP Amsterdam was ready for production (two years ago) – is 
ONAP Dublin used in production? Quo vadis ONAP?

I think that my understanding of a platform just differs. Lets have a look on 
kubernetes: https://kubernetes.io/docs/concepts/overview/components/

There are clear components which each of them is necessary for kubernetes to 
work properly:


  *   Kube-apiserver
  *   Etcd
  *   Kube-scheduler
  *   Kube-controller-manager

If you kill etcd then kubernetes lose its state – apiserver will lose its 
backend and everything going to crash.

Also there is API specification: 
https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.15/

That is the reason why companies like rancher can make its own implementation 
of kubernetes with their own components.

How ONAP compares to this?
--
Best regards,
Petr Ospalý

From: [email protected] [mailto:[email protected]]
Sent: Friday, July 12, 2019 12:52 PM
To: Ospaly Petr <[email protected]>; [email protected]
Subject: Re: [onap-discuss] ONAP as a platform

Hi Petr,

I think you bring-up some interesting points. Let me give my perspective on 
those

1) new platforms are always less optimized, I am not concerned about that, time 
will show. First designs are made for good architecture, not for optimized 
operation, that come later. I have seen too many system designs, which were 
optimized for performance or operation, which failed to deliver.

2) On the platform topic, ONAP has been shown to work for diverse set of use 
cases, so from that perspective I guess it is good enough of a platform.

3) I believe where you have a point is that several projects reused in ONAP 
might need to be well separated out and more used then developed into them (or 
those requirements should be Brough upstream.

4) Also I believe that some choices of tools and components could be revised in 
view of many external projects meanwhile have more matured over time, but they 
were not available or immature at the time the decisions have been taken.

Kind regards,

Marcus

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.

View/Reply Online (#18027): https://lists.onap.org/g/onap-discuss/message/18027
Mute This Topic: https://lists.onap.org/mt/32429722/21656
Group Owner: [email protected]
Unsubscribe: https://lists.onap.org/g/onap-discuss/unsub  
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to