Roughly five weeks have passed since I officially started the work on my PhD. In between I attended a summer school on the convergence of some quite-hype technologies like wireless sensor networks, RFID and peer-to-peer techologies, organized by people from the TU Darmstadt (DVS and KOM). I’m working on a paper about this with a bunch of really nice people. It’s gonne be interesting. I also spent some days in Eindhoven with my group from HU Berlin talking about the research in workflow modelling and analysis and starting a cooperation on common research interests.
In the last three weeks I realized that there is some hap between the idea of having a distributedly designed infrastructure for disaster management and designing applications for these things, which is meant to be the topic of my PhD thesis. The meta-problem seems to be that the these ultra-cool wireless sensor networks won’t be doing much more than sensing and smart routing of data. Not much of a workflow there. But if workflows are meant to be used in a setting with tens or hundreds of thousands of nodes and interconnected devices to assist in disaster management and recovery, where are they going to appear? And assuming we have an answer for that: who or what is going to execute or enact them? The answer to the latter problem is quite likely: some central node with lots of computing power. But we already know how that works (more or less) and there won’t be a new research challenge for me.
Without forgetting about the second question (still hoping for a more challenging answer), I am currently turning to the first question. I might answer it with these new questions:
- What is a workflow in a disaster management system?
- What does it look like?
- What makes it different from other workflows?
- Do we need new formal methods to do so?
- How does self-organization and adaptivity affect these workflows?
One thing that is quite likely to be different and which is a little bit out of focus in the current workflow research are data and resources, which are a crucuial thing in disaster management systems at least. Could be a starting point.