[sg-hub] A new virtual IX in Singapore

[Sanjeev Gupta]

On Fri, Sep 22, 2023 at 1:16 AM Anurag Bhatia <me at anuragbhatia.com> wrote:

> Hi Sanjeev
>
>
> Trying to understand what is the benefit from doing it? I understand your
> cost point about local loop, cross connect etc. But by having a virtual IX
> you are still riding over IP routing of your existing provider (and so do
> others who participate in the virtual IX).
>

Yes, which is one of the reasons I think a local node would be better.  But
not perfect, of course.

There are many things out of my control, even with a local loop; I have no
control over physical cabling, MPLS, etc.  At some point, we do what we
can.  After all, I drive a Korean car, not a Porsche/



> If this is for learning - why not setup private lab. These days you can do
> that over cloud providers easily and for reasonable cost (low cost hourly
> billing).
>

The main issue with a private lab is that it teaches me nothing except copy
and paste.  At least that is all it teaches *me*, YMMV.  Any mistakes I
make, I make on *both* sides, so how would I know?



> If this is for running actual production traffic i.e exchanging live or
> less important but fun traffic, how is that better than say peer to peer
> connection between you and the other network?
>

I may be wrong here, but the problem with bilateral peering is discovery.
Akamai and HE have Peering managers.  I do not.  With a multlateral IX, I
get a foot in the door for free.

SImilar to meeting you to discuss IPv8, compared to a Metup Group.  The
former has better quality, the latter might answer a question I didn't know
I had.


-- 
Sanjeev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.sgnog.net/pipermail/sghub/attachments/20230922/036050fb/attachment.html>