- AI Homework
- Meeting with Prashant
- Candidates Friday
- Spark Emulator
This week was a fairly busy one but not on the research end of things.
The next AI homework was due this week so I spent a significant amount of time on Thursday trying to complete that homework. I unfortunately did not finish the homework. You win some you lose some.
Meeting with Prashant
I met with Prashant to discuss the current state of affairs. We are thinking about writing a negative result paper with MPTCP since it seems like no matter what we do we can not get mptcp to work with some workloads. So we may write a paper classifying workloads and describing why MPTCP can or can not help. We also will include the redundancy benefits of MPTCP as a positive.
Every year we have an open house for new Phd candidates to explore the department and meet professors and current students. It’s a big event and one of my favorite times of the year. This year I’m hosting three students. I spent a lot of this week preparing for this Friday, so research kind of fell by the way side.
In preparation for Friday, Prashant asked me to build a poster to showcase my research, and I do love building posters, so he did not have to ask me twice! I constructed a more general poster on edge clouds as a concept rather than a detailed research poster. I hope to post it soon. This poster will also serve as a good visual for when we have visitors and I need to explain my research to them. I might post it on the wall outside our lab since that seems to be rather barren.
On Monday, I tried running a spark shuffle emulator with larger flight sizes on our cluster. That failed in specatular fashion. It almost immediately crashes some of the machines. I am currently investigating what the cause is and will update this post as necessary.
What’s up for next week?
- New Spark Experiments
- VM migrations
- Tracing with Spark