[dns-operations] DNSViz Service Restoration
casey at deccio.net
Thu Mar 12 05:04:23 UTC 2020
> On Mar 11, 2020, at 2:03 PM, Warren Kumari <warren at kumari.net> wrote:
> On Wed, Mar 11, 2020 at 3:44 PM Matthew Pounsett <matt at dns-oarc.net> wrote:
>> A few weeks ago we made the decision to temporarily put aside the attempt to completely restore the old historical data and instead stand up a new, empty database so that we could get the full featureset online. So, for now there is no access to old tests run prior to the service’s move to OARC, however new tests will be available for review.
>> We’re continuing to work to restore the full historical database; we hope that with the pressure off, and the temptation to cut corners in order to speed up the process removed, we can restart the import from scratch with a slower—but more reliable—approach to recovering those data.
> Excuse my not remembering, but have y'all confirmed that this is
> really worth the faff? What *I* care about is being able to use the
> service *from now on*- going back and seeing the breakage of
> foo.example.net in 2018 is only mildly interesting, but certainly not
> (to me!) worth your time and effort...
Thanks for the perspective. I believe there is value in being able to answer the question: "what did foo.example.net look like at time X?", though it is arguably less important to most users, and the value diminishes over time. Much more important, in my opinion, is understanding trends over time. For example, can we understand what the current pitfalls are and how to improve them, so we can get off NTAs and be more strict in DNSSEC validation adherence? In other words, looking at the aggregate offered by the historical data is much more valuable than looking at individual setups. An understanding of DNS(SEC) deployment behavior can be used to improve tools, processes, and standards.
More information about the dns-operations