[ih] Speaking of layering and gateways

[Jack Haverty]

Thanks, Greg - I agree that there's lots of tools, papers, and 
information within the communities of network operators about today's 
experiences in operating parts of the Internet.  A year or so ago, I 
even gave a keynote talk at one of their conferences - APRICOT2022.

What I was seeking was more of the history of network operations, 
ideally from people who have actually been involved in operating such 
networks.   Not just today's practices, procedures, tools and 
techniques, but also past practices in operating the Internet, and how 
they might have evolved over the 40+ years on Internet operations.

In the days of the later Arpanet, I was "Chief Network Architect" at 
BBN, where we were not only continuing to operate the Arpanet, but also 
implementing the same technology in the DDN and many other networks.  In 
those days (mid-late 1980s) at BBN there was a fairly large group of 
mathematicians who did extensive collection of data about network 
operations, and used it to make decisions such as where to place a new 
node, what capacity lines to interconnect and where to do so, and what 
changes to internal algorithms would have when fully deployed.  They 
also had techniqus to project traffic changes , such as when a new large 
user or application was to be brought online, in order to get necessary 
equipment and circuit changes in place to accommodate the new traffic.

In contrast, I was personally involved in operating internets, first 
when the research Internet was declared to be operational and later when 
working with an international corporate internet.  As I noted earlier, 
most of our decisions were made by instinct and intuition, using 
whatever tools we could find or build and with no "operating manual" to 
be found.   The Internet was and is much more complicated than the 
Arpanet, so the earlier tools and techniques didn't readily apply.  We 
knew how to operate and manage the Arpanet and its clones; we didn't 
know how to do the same for the Internet.

So, for the sake of history, I was hoping that perhaps there were other 
people lurking on this list who may have been involved over the last 3 
decades or so in operating some piece of the Internet or corporate net 
using the same technology, and could comment on their experience - how 
they made decisions, how they collected data to inform such decisions, 
what tools were used, etc.

There's been a lot of discussion on this list of the history of the 
technology as captured in things like RFCs, algorithms, and protocols.   
But there's often a huge gap between documentations and specifications 
and actual practice in the field.  I haven't seen much discussion of the 
operations aspect of the Internet and how it has changed over the years.

Jack Haverty

On 4/16/24 21:37, Greg Skinner via Internet-history wrote:
> On Apr 16, 2024, at 12:30 PM, Jack Haverty via Internet-history<internet-history at elists.isoc.org>  wrote:
>> Perhaps someone has insight into how such decisions are performed today in the Internet, how scientific methods such as OR are used, and how such decision processes have changed over the 50 years of Internet evolution.  Then compare to how its done in other environments such as transportation of goods.  This might make a good topic for some university thesis.
>>
>> Jack Haverty
> I did a bit more googling, this time for [site:nanog.org ''combinatorial optimization''], which turned up a presentation from NANOG 52 on network planning and traffic engineering. [1]   It mentions a tool called pmacct [2] that can take various types of network data that are usable for further analysis.  Perhaps if you searched more of the network operations and measurement web sites, you’d find more of the OR-related work you’re looking for.  It seems to me that a fair amount of this type of work is being done, but if one is not part of those teams, it’s necessary to seek out the individuals belonging to those teams and places where it’s practiced.
>
> --gregbo
>
> [1]https://archive.nanog.org/meetings/nanog52/presentations/Sunday/maghbouleh-bestpractices-tutorial.pdf
> [2]http://www.pmacct.net

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 665 bytes
Desc: OpenPGP digital signature
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20240417/4c47a43e/attachment.asc>