cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Will Stevens <wstev...@cloudops.com>
Subject Re: [DISCUSS] Replacing the VR
Date Tue, 13 Sep 2016 01:33:34 GMT
At this point, I think we are still throwing around ideas.  If you have
links to more options you think we should have on the table, please add
them to the discussion.  :)

*Will STEVENS*
Lead Developer

*CloudOps* *| *Cloud Solutions Experts
420 rue Guy *|* Montreal *|* Quebec *|* H3J 1S6
w cloudops.com *|* tw @CloudOps_

On Mon, Sep 12, 2016 at 6:30 PM, Marty Godsey <marty@gonsource.com> wrote:

> Have we thought of other SDx routers such as pfsense? Its licensed under
> the BSD license and is well maintained.
>
> -----Original Message-----
> From: williamstevens@gmail.com [mailto:williamstevens@gmail.com] On
> Behalf Of Will Stevens
> Sent: Monday, September 12, 2016 6:16 PM
> To: dev@cloudstack.apache.org
> Subject: Re: [DISCUSS] Replacing the VR
>
> You have probably looked into this more than I have Rene.
>
> I am not sure there existed a time when the VR was ever "great".  In my
> eyes, the core ACS dev team should not be building and managing its own
> VR.  I feel like that is a liability because the subset of developers who
> are proficient in networking is quite small.  That means we could be at
> risk of losing the majority of our "experts" with a few people changing
> their $dayjob.  It feels safer to work with an existing technology which
> has its own development community focused on doing that piece well.
> Obviously this has its own drawbacks, but in general, we need the VR
> implementation to be built by dedicated network engineers and not jack of
> all trade developers.  No offense to current company...
>
> I agree with your list of what you would like to see.  Rock solid and not
> over complicated is key.  That being said, if it ONLY handles what ACS
> needs today, then WE have to be the ones to develop any changes.  For
> example; we need IPv6, VXLAN support, etc...  My point is that if we only
> focus on what we need today, then we end up building everything we need for
> the future and I think we end up back where we are now down the road...
>
> I love everything you are saying, just not sure I want us building and
> maintaining it all...
>
> *Will STEVENS*
> Lead Developer
>
> *CloudOps* *| *Cloud Solutions Experts
> 420 rue Guy *|* Montreal *|* Quebec *|* H3J 1S6 w cloudops.com *|* tw
> @CloudOps_
>
> On Mon, Sep 12, 2016 at 5:47 PM, Rene Moser <mail@renemoser.net> wrote:
>
> > Hi
> >
> > On 09/12/2016 10:20 PM, Will Stevens wrote:
> > > *Disclaimer:* This is a thought experiment and should be treated as
> such.
> > > Please weigh in with the good and bad of this idea...
> > >
> > > A couple of us have been discussing the idea of potentially
> > > replacing the ACS VR with the VyOS [1] (Open Source Vyatta VM).
> > > There may be a license issue because I think it is licensed under
> > > GPL, but for the sake of discussion, let's assume we can overcome any
> license issues.
> >
> > VyOS is Debian based, much like the current VR. As long as it is not
> > shipped with CloudStack, all fine.
> >
> > > I have spent some time recently with the VyOS and I have to admit, I
> > > was pretty impressed.  It is simple and intuitive and it gives you a
> > > lot more options for auditing the configuration etc...
> >
> > I had the same "crazy" thoughts when I heard about VyOS the first time.
> >
> > When I looked at VyOS, the release cycle were not very frequent and
> > the current stable release is still based on Debian 6 (EOL [1] since
> > 02.16)
> >
> > However to me, it doesn't matter if it's VyOS or CloudLinux, or
> > another solution.
> >
> > The question is more like what is wrong with the current VR and how
> > can we make the VR great again. Things I would like to see:
> >
> > - VR must have a "clean", programmable, documented, API, supporting
> > batch processing.
> > - VR must be rock solid (minimal shell) state of the art, up to date,
> > but small (Only contain things CloudStack needs, not more)
> > - VR must be scale well (...) and support stateful HA
> > - VR must be easy to upgrade (security) without downtimes.
> >
> > Christmas is soon... ;)
> >
> > René
> >
> > [1] https://www.debian.org/News/2016/20160212
> >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message