Hi Mikael,
Appreciate your inputs. Point taken regarding the buzz words. However need some
clarification about the almost mandatory usage of NAT. I think It is fair to
say that NAT is almost always used where Web servers and Load balancers are
involved. I have a few comments on NAT starting with RFC 4864. RFC 4864:
"Although there are many perceived benefits to Network Address Translation
(NAT), its primary benefit of "amplifying" available address space is not
needed in IPv6."
However NAT is still being used for load balancing (is this reverse
amplification?) with IPv6.
Is NAT being considered dispensable because of the implicit firewall deny/allow
characteristics present due to inside and outside interface configuration ? NAT
configuration is similar to an access rule that only allows particular private
IP to be translated to a public IP or vice versa.
However an implicit deny is present in most Firewalls.
Also NAT local and global definitions appear to be misleading. Taking cue from
your comment about the Cloud being just another computer (host):Inside Network
(or Cloud) and Outside Network (or Cloud) ? Dual Cloud and Dual Network ?This
local and global definition was adopted for IPv6 replacing private and public
IPv4 classification.
So how will NAT be eliminated ?
Is IPv6 striving to eliminate NAT completely or is it only to avoid NAT usage
for conserving address space ?
Thanks & Regards,Vineet Deshpande
On Tuesday, 2 October 2018, 12:17:48 pm GMT+5:30, Mikael Abrahamsson
<[email protected]> wrote:
On Tue, 11 Sep 2018, vineet deshpande wrote:
> Hi,
> I have submitted the updated version on the draft for IP address space
> reclassification. Kindly review.
I started reading your draft:
"An AI based TCP IP model describes how Big Data is centered"
I then proceeded to try to read the rest of the draft. It's completely
full of buzzwords that I do not know what they mean concretely, and it
renders the document impossible for me to understand.
I've been involved in Internet routing for 20 years now, and sections such
as:
" o The Big Data bottleneck lies between different Autonomous
Systems. Due to the implicit life cycle between a Network,
Internet and the Cloud which is time bound and unidirectional,
the Network also needs to grow and scale to the Cloud to reap
the benefits of Cloud such as Datacenter virtualization,
scaling and high availability."
I just don't understand.
" o For such scaling to occur from both ends a Virtual address
space is needed in between. The limitations brought about by
the present classification of IP address space into public,
private and the almost mandatory usage of NAT in the Internet
and Cloud architecture leads to suggest that translating or
mapping of the IP address into another Virtual address space
cannot be avoided."
"almost mandatory usage of NAT". This is being restored to originally
intended state by IPv6 where we do not need NAT anymore (NAT wasn't a
thing until around 2000, before that we didn't do any NAT).
" o The Network cannot be de-classified from the cloud. The
Network needs to grow and scale towards the Cloud."
"The Cloud" is just someone elses computer. Virtualising something doesn't
change anything concretely from it being run on metal.
Please re-write the draft and use concrete terms instead of just a lot of
buzzwords.
--
Mikael Abrahamsson email: [email protected]
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
