[db-wg] What is the RIPE Database in 2021?

Colleagues

There was an interesting presentation in the Address Policy session at
RIPE 82 this morning by Remco. He raised a number of issues about what
exactly is the RIPE Database, what does it do and who does it do it
for? The history of the RIPE Database is quite well known (within the
industry). It developed as a means to assist resource holders and
network operators with 'Internet operational issues'. It also helped
the RIPE NCC, as a registry, to allocate Internet resources fairly and
uniquely as and when needed. It's development and operation was and is
funded by the RIPE NCC membership.

But what is the RIPE Database now, in 2021? What will it be in the
coming years, decades? Who will use it, who will pay for it, who gets
to decide what is in it? The RIPE Database has evolved over the last
20+ years. It is much more now than just a tool for 'Internet
operational issues'. It has a more diverse collection of data
consumers. But is it still locked in the hands of a very small set of
vocal users whose mindset is also locked in the historical context of
the database? Too often in discussions about policies and database
features we hear the phrases "It's not the purpose of the RIPE
Database to do 'abc'" and "It's not the function of the RIPE NCC to do
'xyz'". Are the people saying this RIPE NCC members who are paying for
the database? Is that an issue? I don't know.

Maybe it is time to stop thinking of the RIPE Database only in terms
of 'Internet operational issues'. Whether people have accepted it yet
or not, the RIPE Database is a Public Registry of Internet resources.
It serves a similar important function in society as a Land Registry,
Company Registry, Vehicle Registry, but on a pan national scale. It's
usage goes beyond the historical resource holders, engineers and
network operators. Along with the function of managing the resources
is the importance to the wider public of knowing who is the [owner,
user, registrant, operator] (pick your own word) of an Internet
resource.

To fulfill its role as a Public Registry perhaps it needs more data
not less data. But that has to be the right data. The Database Task
Force (DBTF) is considering a recommendation to make IPv4 assignments
in the RIPE Database optional. That could lead to the complete removal
of 94% of the IPv4 resource detail in the database. Rather than
considering making IPv4 assignments optional, maybe what we should be
considering is making more IPv6 assignment detail mandatory. Do we
need to know who sits behind any of those aggregated blocks? Do we
need to start filling in this detail before we go too far down the
IPv6 road making the backlog difficult to fill in?

Who should make decisions in the future about the purpose, use,
access, content of the RIPE Database? Should it be tied to
(historical) address policy? Or should it be based more on a wider
scope as a Public Registry for the benefit of society? Should the RIPE
NCC members, who are paying for the database, have any priority in
defining it's current and future use? If that becomes an issue should
the RIPE Database be funded independently of the RIPE NCC membership?
These are questions I had expected the DBTF to consider. Unfortunately
I feel the scope of their work has been too narrowly focussed on
current database operational functionality and historical purposes
rather than current and future usage. The first heading in the draft
requirements document is "What is the RIPE Database?". But it only
talks about operational issues. The DBTF clearly stated it is only
reviewing the database functionality. I'm not sure if you can define
the requirements of the functionality without first considering the
full scope of the purpose and usage and the context of the database in
the industry and society. To be fair to the DBTF, their charter also
focuses on database functionality. Maybe there has been a
misunderstanding of the full meaning of 'functionality':

noun (Oxford English Dictionary)

1.the quality of being suited to serve a purpose well; practicality.
"I like the feel and functionality of this bakeware"

2.the range of operations that can be run on a computer or other
electronic system.
"new software with additional functionality"

The draft document seems to focus on the second definition based on a
historical view of the first definition.

Also considering:

The Business Requirement Document (BRD) describes the high-level
business needs whereas the Functional Requirement Document (FRD)
outlines the functions required to fulfill the business need. BRD
answers the question what the business wants to do whereas the FRD
gives an answer to how should it be done.

The draft document from the DBTF is more like the FRD. It outlines the
functions required to fulfill the historical operational needs. We are
still missing the BRD for the RIPE Database in 2021+. We don't have a
clear understanding of what we want the RIPE Database to do and who
for.

Thanks Remco for a very timely presentation. (Although you did still
keep referring to 'operational issues' which keeps the discussion
locked into a historical context.) I look forward to further
discussion on this topic...

cheers
denis
co-chair DB-WG