1 / 18

Consistency Conditions for STM

Consistency Conditions for STM. Sandeep Hans. Agenda. Database Consistency Conditions STM Consistency Conditions A different perspective Consistency with other STM properties. Conclusion. Database Consistency Conditions. Recoverability Avoiding Cascading Aborts Strictness Rigorousness.

oakley
Download Presentation

Consistency Conditions for STM

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Consistency Conditions for STM Sandeep Hans

  2. Agenda • Database Consistency Conditions • STM Consistency Conditions • A different perspective • Consistency with other STM properties. • Conclusion

  3. Database Consistency Conditions • Recoverability • Avoiding Cascading Aborts • Strictness • Rigorousness

  4. Dirty Reads • T1 writes x. • T2 reads x written by T1 . • T2 commits, T1 aborts. w1(x) a T1 r2(x) c2 T2

  5. Recoverability • T1 writes x. • T2 reads x written by T1 . • T1 should commit beforeT2 commits. w1(x) c1 T1 r2(x) c2 T2

  6. Problem with Recoverability • Cascading aborts w1(x) c1 a1 T1 r2(x) c2 a2 T2

  7. Avoiding Cascading Aborts • T1 should commit beforeT2 commits reads x. w1(x) c1 T1 r2(x) c2 T2

  8. Undo Problem • Initially, x=1. • T1 writes x=2. • T2 writes x=3 and commits. • T1 aborts. w1(x=2) a1 T1 w2(x=3) c2 T2

  9. Strictness • No data item is read or overwritten unless the transaction that wrote it has ended. w1(x) c1/a1 T1 r2(x) T2 w3(x) T3

  10. Rigorousness • No data item is read or overwritten unless the transaction that read/wrote it has ended. w1(x) c1/a1 T1 r2(x) c2/a2 T2 w3(x) T3

  11. Landscape RC ACA ST RG [Transactional Information Systems. Gerhard Weikum, Gottfried Vossen.]

  12. STM Conditions • Opacity • Guerraoui and Kapalka[PPoPP’08] • Sequential specification of shared objects. • Virtual World Consistency • Imbs and Raynal [SIROCCO’09] • Causal past of a transaction. • Weakest Reasonable Condition • Doherty , Groves, Luchangco, Moir[REFINE’09]

  13. Database vs. STM • Rigorousness ⊆ Opacity ⊂ Strictness • Rigorousness ⊆ VWC ⊂ Strictness ⊆ ⊆ Are these inclusions strict?

  14. Another Perspective • How do we view the system? • Operation level. • Database consistency conditions. • Recoverability, ACA, Strictness, Rigorousness • Transaction level. • STM consistency conditions. • Opacity, VWC • Is there another level?

  15. Snapshot Isolation • Everyone will have a snapshot of the whole system. • A snapshot must be consistent at every point of time. • Different snapshots need not be consistent with each other. • Updating of snapshots is allowed.

  16. How consistency conditions affect STM properties. • DAP + Invisible Reads + Wait-free + Serializability = Impossible. 1 • We have seen the proof on Wednesday. • How do Opacity/VWC affect such results. • DAP + Wait-free + VWC = possible? • DAP + Wait-free + Opacity = impossible? • 1. Hagit Attiya, Eshcar Hillel, AlessiaMilani. Inherent Limitations on Disjoint-Access Parallel Transactional Memory. SPAA 2009.

  17. Conclusion • Motivation for database and STM consistency conditions is same yet perspectives differ. • STM consistency conditions affect other properties.

  18. Thank You

More Related