Computer Networks - Lecture 1

1
yyfs -A File System for Lots of Tiny Files
Yi Jun, Yuzhe Liu
2
Motivation

• Motivating Applications
• user tend to use the file file in a unusual way
• Generally, the workload that dominated by many
  tiny files references that have the batch
  processing characters

3
Problems

• problems
• tradeoffs between performance and space
  efficiency
• I-node allocation problem
• Problems exist in current G-P file system to
  solve these problems
• Problems exist in existing specially-tailored
  file system to solve these problems

4
Goal to be achieved

• Achieve both high disk space utilization and
  performance
• Do not incur the I-node resource limitation
  problem
• Without compromising the big file reference
  machnisms

5
Design
Figure 1. Layout of structured block
6
Figure 2. Structure of directory entry
7
Figure 3. Blocking compacting
8
Figure 4. System implementation big picture
9
Evaluations

• Workloads
• Simple workloads
• Batch tiny file creation
• Batch tiny file sequential read
• Batch tiny file deletion
• Complex workloads
• Randomly read/delete files
• Randomly batch read/delete file
• metrics
• Number of IO transfer
• Disk space utilization

10
Simple workload-batch file creation
Figure 5. Batch file creation disk efficiency
11
Simple workload-batch file creation-cont.
Figure 6. Batch file creation number of IO
12
Simple workload-Batch file read
Figure 7. Batch file read number of IO
13
Simple workload-Batch file deletion
Figure 8. Batch file delete disk space efficiency
14
Simple workload-Batch file deletion cont.
Figure 9. Batch file delete number of IO
15
Complex workload
Figure 10. Randomly read/delete files
16
Complex workload cont.
Figure 11. Randomly batch read/delete files
17
Drawbacks

• When tiny file is modified to become a large one,
  such modification requires additional operations.
• The storage of directories is not space efficient.

18

• THANK YOU!