5.4 KiB
3rd ACEMS Workshop on Challenges of Data and Control of Networks
I flew to Adelaide to listen and speak at Matt Roughan's 3rd ACEMS Workshop on Challenges of Data and Control of Networks. Matt is an old friend from the AT&T Research Labs family, has co-authored papers with me, and continues research collaboration. So, since he pays, I can go sing for my supper.
The workshop was hosted by The Australian Cyber Security Centre and the Australian Signals Directorate (the Aussie's NSA). So it was data analysis meets security. The ACEMS part was mostly maths folk, so I had tuned my talk to a non-networking audience.
I am waiting for the majority of the talks to be put online.
Goals
I had a number of goals:
-
Learn what I could of the state of the cyber security art from the Aussie gov point of view.
-
The Estonian Gov/Uni cyber security team, lead by an old co-author Olaf Maennel, were sharing the workshop space; and I wanted to learn as much as I could from them.
-
Get maths folk interested in our work discovering actual deployment of routing techniques such as RPKI-based Route Origin Validation, Route Flap Damping, etc. on the real internet.
-
Learn as much as I could about the formal algorithmic folk's current tools for generation and verification of L2/L3, an area in which Matt and his students have been working for a few years.
Successes
Getting current Australian cyber security clue
The Signals Directorate folk have done a lot of work on tools and education for Australian industry on protecting itself. The NSA used to be SigInt (Signals Interception, see Snowden) and ComSec Communications Security i.e. protecting American industry. Sadly, ComSec has pretty much fallen by the wayside in NSA, and the DHS divides its time between abusing children, little bit of ComSec, and ignoring the court system.
The Aussie Signals Directorate, on the other hand, has a lot of literature and tooling to help Aussie industry help itself. The conference host, The Australian Cyber Security Centre, has a lot of material and is very open to sharing.
At the workshops, there were the typical talks on finding weaknesses, tools, etc.
The Estonian Play
Estonia is internetted up the ying yang. One can not do any social or government transaction on paper. They have invested massively in verifiable identity; Olaf's team at Uni Tallinn being a pivot point on which all this is based. Heck, I have been an Estonian E-Citizen for a few years; the identity validation was serious.
All their government databases are interconnected by X-Road, an IP data exchange. Socially, data are much more public than in the American or Japanese cultures; but one can, for example, restrict what provider can see what aspects of one's medical record.
One of their researchers, a German living in Oslo, is actually a psychologist who has added specialization in cyber security. I purposefully cornered him for a long coffee to learn about how we funny monkeys fool ourselves about the limits of our knowledge and abilities to see, anticipate, prevent, ... threats. Then, how can we technologists augment the environment to help compensate for our many human weaknesses?
Collaboration with Uni Adelaide on deployment discovery
To get maths graph theory help, I very intentionally layed out my keynote talk to position the RFD and ROV discovery problems as a graph problem. I had to assume my audience did not know routing, BGP, ...; so half my story was an extremely gentle and simple intro to what BGP they needed to know.
Matt and his ABD, Caitlin Gray, bit, and I am in the middle of socializing the collaboration with the European team with whom I have been working: Pelsser, Wählisch, Schmidt, and Mossig. I can send math if folk are interested.
Maybe
As is common, not everything was easy peasy.
Formal configuration and validation
Matt and his grad students had previously done significant work in formally configuring and validating switching (VLAN) and simple LAN routing to assure paths. Cristel and I had tried to use some of this work in our Cisco-sponsored Work-Flow project. But Matt's key grad student had been sucked into industry. Matt said he had some younger blood who might come up to speed. In the meantime, he did manage to spend some time helping Cristel.
Collaboration with Geoff Huston on ROV deployment discovery
While the graph techniques applicable to RFD deployment are also applicable to detecting ROV deployment; a group of us has been taking a more pragmatic approach and have published previously. But Geoff has his Google Advert platform, where he measures from eyeball browsers. So he has asked if he can use our RPKI-driven ROV testing as data to his eyeball-user platform. I am somewhat optimistic; though eight out of ten attempts to collaborate with Geoff have fizzled.
Concluding
A major reason Australia has good coffee is that the Allies shipped Italian prisoners of war to Australia and they settled. No, I am not kidding.
2020.01.24