[ih] ARPANET pioneer Jack Haverty says the internet was never finished

[Guy Almes]

Hi John,
   This is an interesting thread.
   Your two main points below are at the blurry boundary between 
historical and technical.  Let me put in my two cents' worth.
   Before getting into the points about TOS and multicast, I'll just 
recall several of the strengths of the Internet architecture, including 
the wonderful efficiency and scalability that comes from keeping the 
complicated stuff at the edge (not at the core).  Think of all the times 
when we've rehearsed this in the context of why the TCP/IP Internet 
prevailed over connection-oriented networks.

On 3/2/22 11:03 PM, John Levine via Internet-history wrote:
> It appears that Noel Chiappa via Internet-history <jnc at mercury.lcs.mit.edu> said:
>>    > On Tue, Mar 1, 2022 at 8:46 PM Jack Haverty wrote:
>>
>>    > One that I used in the talk was TOS, i.e., how should routers (and TCPs)
>>    > treat datagrams differently depending on their TOS values.
>>
>>I actually don't think that's that important any more (or multicast either).
>>TOS is only realy important in a network with resource limitations, or very
>>different service levels. We don't have those any more - those limitations
>>have just been engineered away.
> 
> That's not it, they came up against the impenetrable barrier of a
> business model. We understand how to price peering and transit of
> traffic where all packets are the same, but nobody has any idea how
> you do it where some packets are more valuable.
So think about TOS.  The hard problem is not so much the preferential 
packet forwarding (though that does add some unhelpful complexity to 
packet forwarders).  The hard problem is knowing which packets to 
prefer.  And the ISPs would be glad to charge extra for enhanced TOS / 
QOS.  Further, to make a positive statement, in some corporate intranet 
applications, the idea works fine.  But, to do it in the public Internet 
would pull us in the direction of the complexities of the dreaded 
connection-oriented network architectures.  This calls to mind the shock 
some of us Internet engineers had when talking to some telco engineers 
in the late 80s and being told that 30% of the cost of their 
infrastructure was to support billing.

> 
> I never figured out why multicast failed.  It is bizarre that people are dumping
> cable service which has 100 channels multicast to all of the customers in favor
> of point-to-point service where you frequently have a zillion people streaming
> separate copies of the same thing, e.g., a football game.  We fake it with CDNs
> that position servers inside retail networks but really, it's multicast.
The situation with network-layer Multicast is more 
technical/operational.  We have many in our community who are good at 
intra-AS routing and the more difficult but doable us of BGP to set up 
inter-AS routing.  But setting up Multicast routing is much harder and 
less intuitive.  Engineers who are quite good at intra- and inter-AS 
unicast IP routing often find multicast routing very confusing.
The problem isn't the idea of Multicast.
The problem is the enormous hidden costs of doing it at the IP layer.
In contrast, there are many successful examples of applications that do 
Multicast, but at the application layer.  Zoom is an example, but CDNs 
and even the Usenet nntp servers of the 1980s are another.

Coming back to TOS and Multicast together, refraining from burdening the 
router infrastructure of the Internet with solving these problems at the 
IP network layer is part of what allows the Internet to continue to grow 
in its scalability and performance.

But now also coming back to the blend of technical and historical themes 
of the thread, I'd be interested in the thoughts of others.
Is my critique of TOS and network-layer Multicast fair / correct?
And, if so, historically, how did thought evolve on these issues?
	-- Guy
> 
> R's,
> John
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://urldefense.com/v3/__https://elists.isoc.org/mailman/listinfo/internet-history__;!!KwNVnqRv!WVMiL6xyJa6_itoASH6-sGV4p-N2_M6sfqShVAMDVgc8NgiJNQwI6IYSvApCzw$
> 
> .