Data Management Design for Interlaced Magnetic Recording Fenggang Wu - - PowerPoint PPT Presentation

▶
data management design for interlaced magnetic recording
SMART_READER_LITE
LIVE PREVIEW

Data Management Design for Interlaced Magnetic Recording Fenggang Wu - - PowerPoint PPT Presentation

Oct 04, 2023 12 likes •184 views

Data Management Design for Interlaced Magnetic Recording Fenggang Wu , Baoquan Zhang, Zhichao Cao, Hao Wen, Bingzhe Li, Jim Diehl, Guohua Wang*, David H.C. Du University of Minnesota, Twin Cities *South China University of Technology C enter

slide-1
SLIDE 1

Center for Research in Intelligent Storage

Data Management Design for Interlaced Magnetic Recording

Fenggang Wu, Baoquan Zhang, Zhichao Cao, Hao Wen, Bingzhe Li, Jim Diehl, Guohua Wang*, David H.C. Du University of Minnesota, Twin Cities *South China University of Technology

slide-2
SLIDE 2

Center for Research in

Intelligent Storage

Top Tracks Bottom Tracks Conventional Magnetic Recording (CMR) Shingled Magnetic Recording (SMR) Interlaced Magnetic Recording (IMR) Hard Disk Drive IMR: Higher areal data density than CMR, lower write amplification (WA) than SMR.

HDD icon image: https://www.flaticon.com/

slide-3
SLIDE 3

Center for Research in

Intelligent Storage

IMR Tracks Width Laser Power Data Density Data Rate Track Capacity Bottom Tracks wider higher higher(+27%)[1] higher higher Top Tracks narrower lower lower lower lower Updating top tracks has no penalty IMR Updating bottom tracks causes Write Amplification (WA) I/O Performance depends on disk usage, and layout design. Only using bottom tracks when disk is not full may reduce WA.

[1]Granz et. al, 2017

slide-4
SLIDE 4

Center for Research in

Intelligent Storage

  • Adapt to disk usage.
  • Reduce write amplification.
  • Bound memory budget.

The Problem: Data Management Design for IMR

slide-5
SLIDE 5

Center for Research in

Intelligent Storage

  • The problem
  • The solutions

– Baseline design – DM-IMR design

  • The results
  • Future works

Outline

slide-6
SLIDE 6

Center for Research in

Intelligent Storage

Track Group (TG)

OD Top Tracks Bottom Tracks ID Track Group (TG) Track Group (TG) More Track Groups (TGs)

Track Group (TG): an interlaced set of consecutive physical top and bottom tracks. This paper only focus on the data allocation and management within one TG.

slide-7
SLIDE 7

Center for Research in

Intelligent Storage

Three-Phase Baseline

1st Phase 2nd Phase 3rd Phase OD ID Top Tracks Bottom Tracks Track Group (TG) (0~56%) (56~78%) (78~100%)

slide-8
SLIDE 8

Center for Research in

Intelligent Storage

  • Top-Buffer
  • Block-Swap

DM-IMR: Data Management for IMR

slide-9
SLIDE 9

Center for Research in

Intelligent Storage

OD ID Track Group (TG)

Top-Buffer

Top-Buffer Allocated Unallocated The idea: opportunistically buffer bottom-write requests into unallocated top tracks; accumulate multiple updates and write to bottom only once.

slide-10
SLIDE 10

Center for Research in

Intelligent Storage

OD ID Track Group (TG)

Top-Buffer

Top-Buffer Allocated Unallocated 36 78 46 79 lba pba 78 36 Memory Mapping Table 46 79 Design choice: user defines the size budget of the memory table; memory budget determines the max number of tracks Top-Buffer may have. bounded memory budget E.g., If the user bounds the memory table size to be 0.004% of the disk capacity, the max size of the Top-Buffer will be 2% of the disk capacity.

slide-11
SLIDE 11

Center for Research in

Intelligent Storage

OD ID Track Group (TG)

Top-Buffer

Top-Buffer Allocated Unallocated Top-Buffer capacity also depends on available unallocated top tracks. Problem:

  • Extremely small Top-Buffer brings little benefit.
  • Top-Buffer cannot function when usage=100%.

X1 Y1 X2 Y2 lba pba Memory Mapping Table bounded memory budget X3 Y3 X4 Y4 X5 Y5

slide-12
SLIDE 12

Center for Research in

Intelligent Storage

Block-Swap

Top-Buffer Block-Swap ID OD Top-Buffer Allocated Track Group (TG) 36 78 46 79 76 24 27 80 24 76 lba pba The idea: progressively swap hot data in bottom tracks with cold data in top tracks. 76 46 78 79 24 36 80 27 Memory Mapping Table bounded memory budget Design choice: Top-Buffer and Block-Swap share the memory budget; Block-Swap will kick in when Top-Buffer cannot fully use the mapping table (i.e. usage is high).

slide-13
SLIDE 13

Center for Research in

Intelligent Storage

DM-IMR: Put it together

Space utilizations (%)

56% 78% 100% (0~56%) Top-Buffer: at most 2% of the whole space

Bottom Update Scheme

Top-Buffer Block-Swap (56~78%) (78~100%) 98% In-Place (more design details in paper)

slide-14
SLIDE 14

Center for Research in

Intelligent Storage

  • IMR Sim
  • MSR Cambridge Trace Replay
  • Competing Schemes

Evaluation

Three-Phase Baseline Buffer-Only

Space utilizations (%)

56% 78% 100%

Bottom Update Scheme

In-Place Top-Buffer 98%

Space utilizations (%)

56% 78% 100%

Bottom Update Scheme

In-Place 98%

slide-15
SLIDE 15

Center for Research in

Intelligent Storage

Average Throughput with Varying Usage

  • Buffer-Only and DM-IMR both can increase throughput.
  • DM-IMR outperforms Buffer-Only after 98% because Block-Swap starts

to kick in.

DM-IMR

Block-Swap kicks in Higher = Better

Space utilizations (%)

56% 78% 100%

Bottom Update Scheme

Top-Buffer Block-Swap 98% In-Place

More results in the paper

slide-16
SLIDE 16

Center for Research in

Intelligent Storage

  • Problem: data management for IMR.
  • Two approaches are proposed:

– Three-Phase baseline – DM-IMR, which uses Top-Buffer and Block-Swap to improve from the Three-Phase baseline.

  • Results show DM-IMR can increase throughput and reduce write

amplification.

  • Future work: space manager design for TGs, eviction algorithms of Top-

Buffer and Block-Swap, computation optimization, etc.

Summary

slide-17
SLIDE 17

Center for Research in Intelligent Storage

Data Management Design for Interlaced Magnetic Recording

Thank you! Comments/Questions?