Hi all,

In the current Internet, an ICP (e.g., google) could scatter a set of
servers sharing the same FQDN (e.g., www.google.com ) worldwide or
nationwide to provide load-balancing services in an anycast-like way. For
example, if a client performs a DNS query for the A or AAAA records of ”
www.google.com”, the DNS server could return the address of a server which
is one of the nearest servers to the client, in the DNS response. Of course,
the DNS server could also return all available A or AAAA records, then the
client will choose one according to some policy. In this case, the FQDN
plays a role of service ID, rather than host ID.

If I understood ILNP correctly, each ILNP host needs a globally unique FQDN.
That’s to say, the FQDN can not be used as a service ID to represent a set
of ILNP servers which are scattered in the Internet. As a result, the above
load-balancing service through DNS is not available anymore in the ILNP
architecture. Maybe another indirection from one FQDN (representing a
Service ID) to another one (representing a Host ID) should be developed in
order to support the above service.


> -----邮件原件-----
> 发件人: Xu Xiaohu [mailto:x...@huawei.com]
> 发送时间: 2010年1月28日 16:51
> 收件人: 'Tony Li'
> 抄送: 'RRG'
> 主题: re: [rrg] A concern with ILNP//re: critique of RANGI
> > -----邮件原件-----
> > 发件人: Tony Li [mailto:tony...@tony.li]
> > 发送时间: 2010年1月28日 16:15
> > 收件人: Xu Xiaohu
> > 抄送: 'RRG'
> > 主题: Re: [rrg] A concern with ILNP//re: critique of RANGI
> >
> > Xu Xiaohu wrote:
> >
> > > OK, I got it. My point is the above scenario should be considered in a
> > > clean-slate architecture.
> >
> >
> > Not necessarily.  Some folks don't feel that it's necessary (or
> > beneficial) to warp the architecture around the legacy host.  Things do
> > change.  We no longer need to be concerned with our IMP number, for
> > example.  ;-)
> How many people used IMP, and how many people use current Internet? The
> difference in scale may cause us to make totally different choices. ;)
> Xiaohu
> > Yes, we could have a band-aid architecture.  But then we have to live
> > with those band-aids.  Forever.  No thank you.
> >
> >
> > > Then each ILNP+ host should be assigned a globally unique home address
> > > for communication with legacy hosts ;)
> >
> >
> > There's no such thing as ILNP+.  If someone wants to use an ILNP host
> > with Mobile IP to interact with legacy hosts, then yes, it will need to
> > act like a legacy host and have a home address.
> >
> >
> > > Not better, but worse in some cases (e.g., the above scenario), IMHO.
> >
> >
> > You haven't shown that.  All you've done is to restate that legacy hosts
> > can't take advantage of ILNP.  This was stipulated up front.  Asked and
> > answered.  Let's move on.
> >
> > Tony

rrg mailing list

Reply via email to