[metrics-team] Updated roadmap from October 2017 to September 2018

Karsten Loesing karsten at torproject.org
Thu Apr 19 19:49:20 UTC 2018

Hi team,

as discussed last week in Aberdeen and also team-internally in the past
few days, here are the updates to our roadmap until September 2018.

The reason why I'm sending this update here rather than publish them on
the wiki page are: we can link to this post rather than flood the wiki
page with a long explanation, and people interested in the metrics team
will notice the update.

To give others some more context, we already have a roadmap for the time
from October 2017 to September 2018.


However, there are a few reasons why we're updating this 12-month
roadmap now.

First, we're not growing the team in July from 2 to 3 full-time
developers after all: we are going to add irl full time, but we're also
losing iwakeh, which may or may not be temporary.

Second, the sponsored goals require quite some more effort than we had
expected in the beginning.

And third, we received new requests from other teams at the last Tor
meeting which we should consider.

Most of our original goals will remain unchanged, either because they
are already completed or because we think we'll simply have to keep them:

 2. Switch ExoneraTor and metrics-web from Tomcat to embedded Jetty, and
harmonize directory structures (Code; M; 100% done)
 4. Make sure that each service has at least two operators (Operation;
M; 20% done)
 5. Add web server logs as new data source (Sources; M; 100% done)
 6. Provide metrics timeline events as both a table on Tor Metrics pages
and as an RSS/Atom feed that is also syndicated via Twitter to increase
community engagement (News; M; 60% done)
 7. Make Atlas part of the Tor Metrics website (Services; M; 90% done)
 8. Make Compass part of Atlas in order to provide aggregated search
results and details pages on Atlas and to finally shut down Compass
(Services; M; 100% done)
 10. Add OONI graphs to Tor Metrics based on pre-aggregated data from
vanilla Tor test and bridge reachability test (Visualizations; M; 40% done)
 11. Add IPv6 relay graphs to Tor Metrics based on existing descriptor
contents (Visualizations; M; 100% done)
 14. Put external research data on the website, including Rob's
phantomtrain OnionPerf data and Yixin's BGP monitoring data (Research;
M; 30% done)
 17. Help the Core Tor and Tor Browser teams with ongoing measurements
(Sources; M; 50% done)

For some goals we're changing our estimation of overall effort. The
first is a goal that we estimated to be an L task which turned out to be
an M task:

 16. Support PrivCount development and integrate its data as successor
to statistics reported by relays and bridges (Sources; M; 50% done)

Another goal that we'd like to adapt is the first deliverable for
Sponsor 13. We'd like to ask for an extension of 3 months for a total of
15 months of which 12 months (4/5) would fall into the planned period.
The goal would be to reach 80% by September 2018, not 100% as is the
case with most other goals.

 18. Document the Tor Metrics system/pipeline, including explicitly
writing down our requirements (Sponsor 13) (About; 4/5 L; 5% of 80% done)

The same applies to the second deliverable for Sponsor 13, which also
turned out to be more time-consuming as originally expected. That's why
we're changing our estimation to 2 L of which we're planning to complete
8/5 L or 80% by September 2018.

 13. Document aggregated statistics files better (Sponsor 13) (Research;
8/5 L; 35% of 80% done)

From talking to other teams at the last Tor meeting, we think we'll have
to add two new goals to have the roadmap reflect what we're going to
work on.

 19. Extend Onionoo to use JNDI for rDNS lookups (Services; M; 0% done)
 20. Add a few more performance graphs based on existing OnionPerf data
(Visualizations; M; 0% done)

Finally, we need to remove some goals from the roadmap to make it a
realistic plan. As a result we're going to stop working on the following
goals. Progress made up to this point will be counted for the burndown
chart and capacity planning, but remaining effort will be set to 0%, as
if these tasks were done now. The following were all hard decisions, but
we can always resume working on them in six months from now.

 1. Update all code to use Java 8 features (Code; M; 30% of 30% done,
 3. Deploy better notification system for operational issues (Operation;
M; 30% of 30% done, deferred)
 9. Resolve ExoneraTor database slowness (Services; M; 25% of 25% done,
 12. Rename Tor Metrics components to make them easier to understand for
people outside of the team (About; M; 45% of 45% done, deferred)
 15. Re-process bridge descriptor tarballs to retain “contact” lines in
server descriptors and “fingerprint” lines in statuses (Sources; M; 25%
of 25% done, deferred)

The goals above sum up to 24.15 M tasks. In the first six months of the
roadmap period we finished 11.95 M, which leaves 12.20 M as remaining
work for the second six months.

The new team capacity is 24 M, down from 27 M as originally planned.
That's 12 M for the first six months and 12 M for the second six months.

Overall, we're only slightly behind our goals in this updated roadmap.
That is, assuming the Sponsor 13 extension goes through and we don't add
anything else. In the latter case we already considered deferring
remaining work on deliverables 6, 10, and 14, which would free another
1.7 M tasks at this point.

I'm going to update the wiki page tomorrow to reflect these changes.

All the best,

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 528 bytes
Desc: OpenPGP digital signature
URL: <http://lists.torproject.org/pipermail/metrics-team/attachments/20180419/bf8bfed6/attachment.sig>

More information about the metrics-team mailing list