DOC PREVIEW
U of U CS 7810 - Disk Data Layer & Other Issues

This preview shows page 1-2-20-21 out of 21 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 21 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 21 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 21 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 21 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 21 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Page 1 1 CS7810 School of Computing University of Utah Disk Data Layer & Other Issues Reference: “Memory Systems: Cache, DRAM, Disk” – Jacob, Ng, & Wang, Ch. 18 “Failure Trends in a Large Disk Drive Population” – FAST07 – Pinheiro, Weber, Barroso, Google 1955: IBM RAMAC 305 Today: Hitachi MicroDrive 2 CS7810 School of Computing University of Utah Fixed Size Blocks • Ideal case  contiguous placement » one seek and rotational latency hit • sector reordering in buffer reduces the rotational impact » hard to do for large files • Block size choice – fixed vs. variable  common memory theme » page allocation to physical memory • Fixed size blocks  + one unit of allocation – easy map to sectors » - can’t find enough contiguous ones to meet need • fix with compaction/GC but this takes time » - internal fragmentation • last block will always only be partially full – not a big deal w/ today’s densities unless lots of very small files  common: map file to non-contiguous blocks » trades file access time for reduced GCPage 2 3 CS7810 School of Computing University of Utah Variable Size Blocks • + No internal fragmentation • Used in early systems  primarily due to small file sizes and crappy density • Several problems  large file writes » not enough contiguous free space • GC must happen more often  accounting woes » free space entries must be done at a finer grain size • increased level of meta-data » count field must be associated with each record • or held in metadata 4 CS7810 School of Computing University of Utah Sectors 1 • Today’s choice  sectors = fixed sized blocks  specified by interface standards » consumer: SATA » and commercial: SCSI, SAS, and FC • Anatomy of a sector  gap » allow time for read and write heads to access the sector » writes – jitter buffer to prevent adjacent sector contamination » also compensates for some drift in clock  preamble (a.k.a. sync field) » ~ 10 bytes » establish frequency and amplitude of recorded signals • adjust the PLL and AGC circuitsPage 3 5 CS7810 School of Computing University of Utah Sectors 2 • All are vendor specific choices  data address mark (a.k.a. data sync field) » a few bytes » special pattern indicates end of preamble and beginning of actual data  data » 512 bytes RLL encoded into ~544 bytes  CRC » correct one-time data errors  ECC » ~40 bytes: burst ECC required • both hard and soft error’s take out an area 6 CS7810 School of Computing University of Utah Sector Size • Tradition  512 bytes » BIOS, drivers, and file system coded for 512B  problems » as areal density increases • burst errors get bigger • both SER and HER probabilities increase » hence bigger ECC and CRC fields required • efficiency of data to overhead fields goes up • Fix w/ bigger sectors in 2005  4 KB & 512B allowed » MS Vista written w/ 4KB in mind • Internal larger sector  take pieces of it to make it look 512B at the interfacePage 4 7 CS7810 School of Computing University of Utah Tracks and Cylinders • Track options  concentric rings – used for disks  spiral – used for continuous media, CD, DVD, vinyl records • Cylinders  vertical set of same tracks on different surfaces » modern precision & variance  not purely aligned » hence switching heads entails minor track center correction • max transition ΔV indicates center • Numbering  Sectors: 1 to N for each track  Tracks: outer track is 0, inner track is m  Negative tracks (-1 … -n): even more outer » reserved cylinders to hold non-user data • defect maps • address maps • etc. 8 CS7810 School of Computing University of Utah Address Mapping • External access via sector ID (integer)  essentially a logical block address (LBA) • Internally mapped to a physical block address (PBA)  a.k.a. absolute block address (ABA)  in reality a CHS index » cylinder, head, sector • Next ABA options (n-1 ends track, n starts a new track)  cylinder mode » n is on a different surface • minimizes seek  track mode » stay w/ existing head • minimized head electronics switching delay » turns out today: track mode wins • primarily because tracks are not perfectly vertically aligned in cylinder • hence seek anyway but which directionPage 5 9 CS7810 School of Computing University of Utah Jaggy Cylinders High TPI density increases variance 10 CS7810 School of Computing University of Utah Serpentine Mapping • Problem  stay on same surface as long as possible  AND stay with out to in numbering  long seek to change surface » no need to be stuck with this but devices tend to anywayPage 6 11 CS7810 School of Computing University of Utah Skewing • Track mode – same surface issue  can’t seek in gap time » add one whole rev. time to next access  so stagger the #1 sector (track skew) • Same game can be played in cylinder mode  cylinder skew 12 CS7810 School of Computing University of Utah Variable Density Recording • Each bit placed on a radial  fixed # of sectors/track  bpiMAX on innermost track » cons • wasted capacity increases as you move to outer tracks » pros • motor rpm stays constant • bit rate at the heads is always the same – simplifies timing compliance  capacity » = tpi x bpi x Π x (OD2-ID2)/4  today’s high density disks » too much lost capacity » higher $/bit  bad ideaPage 7 13 CS7810 School of Computing University of Utah Fixed Density Variable RPM • Pro’s  no wasted capacity  timing per bit constant • BIG NEGATIVE  change RPM based on track address » RPM takes 100’s ms to stabilize » .5 rotational latency varies by track • 3@ID – 6@OD ms • > 10x penalty due to the motor • Also a bad idea 14 CS7810 School of Computing University of Utah Zoned Bit Recording (ZBR) • Divide collections of tracks into zones  fixed density and RPM  vary # sectors per zone  fixed # sectors/zone » causes bit rate


View Full Document

U of U CS 7810 - Disk Data Layer & Other Issues

Download Disk Data Layer & Other Issues
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Disk Data Layer & Other Issues and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Disk Data Layer & Other Issues 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?