1. Distributed MapReduce Team B Presented by: Christian Bryan Matthew Dailey Greg Opperman Nate Piper Brett Ponsler Samuel Song Alex Ostapenko Keilin Bickar

2. Introduction

3. Functional Languages

4. What makes MapReduce Special?  Map function  Lisp – McCarthy et al in 1958  Reduce function  (paper example) = Summing up occurrences  The combination?  Behind the scenes action  One user to n computers, where the only insight into n is the speed at which computation is completed.

5. Example of an abstraction (define appendEven (lambda (x) (cond ((empty? x) empty) (else (begin (cond ((= 0 (remainder (car x) 2)) (cons (car x) (appendEven (cdr x)))) (else (cons (* 2 (car x)) (appendEven (cdr x)) )))))))) (define appendEvenMap (lambda (x) (cond ((= 0 (remainder x 2)) x) (else (* 2 x))))) (appendEven myL) (map appendEvenMap myL) (list 2 2 6 4 10 6 14 8 18 10 22 12)

6. How many computers is the user connected to? MapReduce hides this implementation decision. SCALABLE???

7. Goals of Distributed System Transparency Scalable More fault tolerant than standalone system Monotonicity – Can’t retract statements Which computer is correct? Many points of failure Problems when scaling

8. Naturally Distributable

9. Why? The 'map' and 'reduce' functions themselves. 'map' takes in a function and a set of data. That set of data is partitioned and ready to go. Function + Data = Convenient

10. More... 'reduce' is less convenient. Takes in an operation and a dataset. GFS helps out alot.

11. Distributing Map and Reduce User writes Map function o (k1,v1) → list(k2,v2) Next, user rights Reduce program o (k2,list(v2)) → list(v2) Specification file defines inputs, outputs, and tuning parameters o Passed to MapReduce function MapReduce library handles the rest!

13. Productivity Improvements Programmers no longer have to program for the network Simplified library to make a program distributed, can be reused Can focus on problem instead of distributed implementation of it Quote from Google: "Fun to use" o Programmers having fun are more productive

14. MapReduce Performance

15. Measured Performance ~1800 2GHz processors with 4Gb of RAM used First test task – search through ~1Tb of data for a particular pattern Second test task – sort ~1Tb of data

16. Test 1 (searching) Input split into 64Mb pieces Machines assigned until all are working @55sec Sources of delay: startup, opening files, locality optimization

17. Test 2 (sorting)

18. Criticism from Database Systems Community Very old concepts used Poor implementation (indices) Limited set of features (idea of views)

19. Fault Tolerance

20. Worker Failure Master pings workers periodically Worker “fails” if it does not respond within a certain amount of time. All map tasks completed or in progress by worker are reset to idle state Eligible for rescheduling

21. Worker Failure Completed Reduce tasks are not reset because their output is stored in a global file system and not locally on the Failed Machine. All workers are notified of the changes in workers Resilient to large scale worker failure.

22. Master Failure Periodic checkpoints Upon Failure: a new copy starts from last checkpoint Failure of master is unlikely Current implementation aborts upon Master Failure

24. Google Cluster Configuration Large clusters of commodity PCs connected together with switched Ethernet Typically dual-processor x86 processors running Linux, 2-4 GB of memory Inexpensive IDE disks attached directly to individual machines Commodity networking hardware is used. Typically either 100 megabits/second or 1 gigabit/second at the machine level

25. Google Cluster Operation Users submit jobs to a scheduling system. Each job consists of a set of tasks, and is mapped by the scheduler to a set of available machines within a cluster. A distributed file system (GFS) is used to manage the data stored on the disks. Uses replication to provide availability and reliability on top of unreliable hardware.

26. Networking

27. Cost Efficiency April 2004, Google spent about $250 million on hardware equipment o includes other equipment than CPUs such as routers and firewalls o Approximately  63, 272 machines  126,554 CPUs  253, 088 GHz of processing power  126,544 Gb of RAM  5,062 TB of Hard Drive Space  About 253 teraflops (trillion floating point operations per second)

28. Cost Efficiency January 2005, Japan's NEC's Earth Simulator supercomputer o $250 million o 41 teraflops Much more expensive compared to a large cluster of personal computers

29. Cost Efficiency 2003, Virgina Tech used 1,100 Apple computers o cost $5 million o 10 teraflops o 3rd most powerful at the time o supercomputer would have cost much more

30. Cost Efficiency Disadvantages o deal with network bandwidth o constantly monitor for hardware failure

31. Conclusion

32. Questions?