[Presentation] Report meeting: still a lot to do!

[fdemoor]

Introduction

• Add Map-Reduce introduction, explain how it works, how tuples are distributed from map tasks to reduce tasks
• Introduce data skew

Data Skew Challenges

• Causes and consequences of data skew
• Why important research subject

Handling Data Skew

• Presented algorithms focus on reduce skew
• Two types of solutions:
  • LEEN and FP / DF for partitioning skew
  • SkewTune for reduce complexity
• Present briefly algorithms
• Small overview at the beginning is good, but must be brief, with relevant keywords
• Last section about related work / other approached is very good idea, reduce other subsection if needed

Conclusion

• Recall problem and why interesting
• Recall solutions
• Open on perspectives, possible improvements (grammar based prediction for text datasets?)

General remarks

• Fair partition of speak time
• Get to work
• D-Day: 5 & 6 April

[fdemoor]

done;