Survey Paper - UCE15Asynchronous Replication of  Databases in Large Scale                 State Of The Art       Masters D...
Survey Paper - UCE15                    AgendaIntroductionConsistency CriteriaAsynchronous ReplicationReplication in Large...
Survey Paper - UCE15     IntroductionMasters Degree in Computer Engineering          University of Minho
Survey Paper - UCE15               IntroductionWhy replicate data?          Masters Degree in Computer Engineering        ...
Survey Paper - UCE15                IntroductionWhy replicate data?   Fault-Tolerance           Masters Degree in Computer...
Survey Paper - UCE15                 IntroductionWhy replicate data?   Fault-Tolerance   Performance           Masters Deg...
Survey Paper - UCE15                 IntroductionWhy replicate data?   Fault-Tolerance   Performance           Masters Deg...
Survey Paper - UCE15                  IntroductionWhy replicate data?   Fault-Tolerance   PerformanceHowever...           ...
Survey Paper - UCE15                  IntroductionWhy replicate data?   Fault-Tolerance   PerformanceHowever...   Introduc...
Survey Paper - UCE15                  IntroductionWhy replicate data?   Fault-Tolerance   PerformanceHowever...   Introduc...
Survey Paper - UCE15     IntroductionMasters Degree in Computer Engineering          University of Minho
Survey Paper - UCE15                IntroductionClassification of replication protocols:           Masters Degree in Comput...
Survey Paper - UCE15                IntroductionClassification of replication protocols:   When can updates be performed:  ...
Survey Paper - UCE15                IntroductionClassification of replication protocols:   When can updates be performed:  ...
Survey Paper - UCE15                IntroductionClassification of replication protocols:   When can updates be performed:  ...
Survey Paper - UCE15                IntroductionClassification of replication protocols:   When can updates be performed:  ...
Survey Paper - UCE15Consistency Criteria Masters Degree in Computer Engineering           University of Minho
Survey Paper - UCE15              Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranc...
Survey Paper - UCE15              Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranc...
Survey Paper - UCE15              Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranc...
Survey Paper - UCE15              Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranc...
Survey Paper - UCE15                Consistency CriteriaKey issue, since replication is a solution to achieve fault-tolera...
Survey Paper - UCE15             Consistency CriteriaOther Criteria:   Increasing popularity of Snapshot Isolation (SI)   ...
Survey Paper - UCE15Asynchronous Replication    Masters Degree in Computer Engineering              University of Minho
Survey Paper - UCE15        Asynchronous ReplicationLazy schemes update replicas using separate transactions              ...
Survey Paper - UCE15         Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the co...
Survey Paper - UCE15         Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the co...
Survey Paper - UCE15         Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the co...
Survey Paper - UCE15         Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the co...
Survey Paper - UCE15Replication in Large Scale and Data             Freshness        Masters Degree in Computer Engineerin...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15 Replication in Large Scale and Data              Freshness  Current techniques have attained some deg...
Survey Paper - UCE15Replication in Large Scale and Data             Freshness        Masters Degree in Computer Engineerin...
Survey Paper - UCE15  Replication in Large Scale and Data               Freshness Although serializability is guaranteed o...
Survey Paper - UCE15Replication in Large Scale and Data             Freshness Transactions may see stale data They may be ...
Survey Paper - UCE15           MySQLMasters Degree in Computer Engineering          University of Minho
Survey Paper - UCE15                          MySQLMySQL implements asynchronous master-slave replication               Ma...
Survey Paper - UCE15                             MySQL MySQL implements asynchronous master-slave replication Uses the pri...
Survey Paper - UCE15                             MySQL MySQL implements asynchronous master-slave replication Uses the pri...
Survey Paper - UCE15                             MySQL MySQL implements asynchronous master-slave replication Uses the pri...
Survey Paper - UCE15                           MySQLReplication topologies:                Masters Degree in Computer Engi...
Survey Paper - UCE15                                        MySQL   Replication topologies:Master and Multiple Slaves     ...
Survey Paper - UCE15                                        MySQL   Replication topologies:Master and Multiple Slaves     ...
Survey Paper - UCE15                                        MySQL   Replication topologies:Master and Multiple Slaves     ...
Survey Paper - UCE15   ConclusionsMasters Degree in Computer Engineering          University of Minho
Survey Paper - UCE15                        ConclusionsEager protocols are not appropriate for large scale systems        ...
Survey Paper - UCE15                        ConclusionsEager protocols are not appropriate for large scale systemsLazy pro...
Survey Paper - UCE15                         ConclusionsEager protocols are not appropriate for large scale systemsLazy pr...
Survey Paper - UCE15                         ConclusionsEager protocols are not appropriate for large scale systemsLazy pr...
Survey Paper - UCE15                         ConclusionsEager protocols are not appropriate for large scale systemsLazy pr...
Survey Paper - UCE15                         ConclusionsEager protocols are not appropriate for large scale systemsLazy pr...
Survey Paper - UCE15                         ConclusionsEager protocols are not appropriate for large scale systemsLazy pr...
Upcoming SlideShare
Loading in …5
×

Asynchronous Replication of Databases

1,232 views

Published on

Presentation for survey paper about asynchronous replication of databases

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,232
On SlideShare
0
From Embeds
0
Number of Embeds
16
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Asynchronous Replication of Databases

  1. 1. Survey Paper - UCE15Asynchronous Replication of Databases in Large Scale State Of The Art Masters Degree in Computer Engineering University of Minho
  2. 2. Survey Paper - UCE15 AgendaIntroductionConsistency CriteriaAsynchronous ReplicationReplication in Large Scale and Data FreshnessMySQLConclusions Masters Degree in Computer Engineering University of Minho
  3. 3. Survey Paper - UCE15 IntroductionMasters Degree in Computer Engineering University of Minho
  4. 4. Survey Paper - UCE15 IntroductionWhy replicate data? Masters Degree in Computer Engineering University of Minho
  5. 5. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance Masters Degree in Computer Engineering University of Minho
  6. 6. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance Performance Masters Degree in Computer Engineering University of Minho
  7. 7. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance Performance Masters Degree in Computer Engineering University of Minho
  8. 8. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance PerformanceHowever... Masters Degree in Computer Engineering University of Minho
  9. 9. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance PerformanceHowever... Introduces a constant trade-off between consistency and performance Masters Degree in Computer Engineering University of Minho
  10. 10. Survey Paper - UCE15 IntroductionWhy replicate data? Fault-Tolerance PerformanceHowever... Introduces a constant trade-off between consistency and performance Need to use adequate replication mechanisms / protocols Masters Degree in Computer Engineering University of Minho
  11. 11. Survey Paper - UCE15 IntroductionMasters Degree in Computer Engineering University of Minho
  12. 12. Survey Paper - UCE15 IntroductionClassification of replication protocols: Masters Degree in Computer Engineering University of Minho
  13. 13. Survey Paper - UCE15 IntroductionClassification of replication protocols: When can updates be performed: Masters Degree in Computer Engineering University of Minho
  14. 14. Survey Paper - UCE15 IntroductionClassification of replication protocols: When can updates be performed: Who can perform updates: Masters Degree in Computer Engineering University of Minho
  15. 15. Survey Paper - UCE15 IntroductionClassification of replication protocols: When can updates be performed: - Lazy (also know as asynchronous) - Eager (also know as synchronous) Who can perform updates: Masters Degree in Computer Engineering University of Minho
  16. 16. Survey Paper - UCE15 IntroductionClassification of replication protocols: When can updates be performed: - Lazy (also know as asynchronous) - Eager (also know as synchronous) Who can perform updates: - Primary copy - Update-everywhere Masters Degree in Computer Engineering University of Minho
  17. 17. Survey Paper - UCE15Consistency Criteria Masters Degree in Computer Engineering University of Minho
  18. 18. Survey Paper - UCE15 Consistency CriteriaKey issue, since replication is a solution to achieve fault-tolerance Masters Degree in Computer Engineering University of Minho
  19. 19. Survey Paper - UCE15 Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranceCorrect behaviour in a replicated system must ensure linearizability(also know as one copy equivalence): Masters Degree in Computer Engineering University of Minho
  20. 20. Survey Paper - UCE15 Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranceCorrect behaviour in a replicated system must ensure linearizability(also know as one copy equivalence): ‣ Gives the illusion that a replicated database system is single Masters Degree in Computer Engineering University of Minho
  21. 21. Survey Paper - UCE15 Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranceCorrect behaviour in a replicated system must ensure linearizability(also know as one copy equivalence): ‣ Gives the illusion that a replicated database system is single ‣ Maintain order Masters Degree in Computer Engineering University of Minho
  22. 22. Survey Paper - UCE15 Consistency CriteriaKey issue, since replication is a solution to achieve fault-toleranceCorrect behaviour in a replicated system must ensure linearizability(also know as one copy equivalence): ‣ Gives the illusion that a replicated database system is single ‣ Maintain order ‣ Atomicity Masters Degree in Computer Engineering University of Minho
  23. 23. Survey Paper - UCE15 Consistency CriteriaOther Criteria: Increasing popularity of Snapshot Isolation (SI) Strong Serializability Strong Session 1SR Masters Degree in Computer Engineering University of Minho
  24. 24. Survey Paper - UCE15Asynchronous Replication Masters Degree in Computer Engineering University of Minho
  25. 25. Survey Paper - UCE15 Asynchronous ReplicationLazy schemes update replicas using separate transactions Masters Degree in Computer Engineering University of Minho
  26. 26. Survey Paper - UCE15 Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the complexity and performance of eager replication, there isa wide spectrum of lazy schemes Masters Degree in Computer Engineering University of Minho
  27. 27. Survey Paper - UCE15 Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the complexity and performance of eager replication, there isa wide spectrum of lazy schemesLazy Replication Models: Masters Degree in Computer Engineering University of Minho
  28. 28. Survey Paper - UCE15 Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the complexity and performance of eager replication, there isa wide spectrum of lazy schemesLazy Replication Models: Primary copy Masters Degree in Computer Engineering University of Minho
  29. 29. Survey Paper - UCE15 Asynchronous ReplicationLazy schemes update replicas using separate transactionsDue to the complexity and performance of eager replication, there isa wide spectrum of lazy schemesLazy Replication Models: Primary copy Update-everywhere Masters Degree in Computer Engineering University of Minho
  30. 30. Survey Paper - UCE15Replication in Large Scale and Data Freshness Masters Degree in Computer Engineering University of Minho
  31. 31. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Masters Degree in Computer Engineering University of Minho
  32. 32. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Most solutions adopt a full replication model Masters Degree in Computer Engineering University of Minho
  33. 33. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Most solutions adopt a full replication model ➡ Coordination overhead Masters Degree in Computer Engineering University of Minho
  34. 34. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Most solutions adopt a full replication model ➡ Coordination overhead Most solutions rely on 1-copy-serializability Masters Degree in Computer Engineering University of Minho
  35. 35. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Most solutions adopt a full replication model ➡ Coordination overhead Most solutions rely on 1-copy-serializability ➡ Limits concurrency Masters Degree in Computer Engineering University of Minho
  36. 36. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Current techniques have attained some degree of scalability,however there are two main limitations: Most solutions adopt a full replication model ➡ Coordination overhead Most solutions rely on 1-copy-serializability ➡ Limits concurrency ➡ Limits scalability of the system Masters Degree in Computer Engineering University of Minho
  37. 37. Survey Paper - UCE15Replication in Large Scale and Data Freshness Masters Degree in Computer Engineering University of Minho
  38. 38. Survey Paper - UCE15 Replication in Large Scale and Data Freshness Although serializability is guaranteed on lazy replication system withconcurrency control techniques and consistency criterion, thesetechniques do not provide data freshness guarantees: Masters Degree in Computer Engineering University of Minho
  39. 39. Survey Paper - UCE15Replication in Large Scale and Data Freshness Transactions may see stale data They may be serialized in an order different from the one inwhich they were submitted Masters Degree in Computer Engineering University of Minho
  40. 40. Survey Paper - UCE15 MySQLMasters Degree in Computer Engineering University of Minho
  41. 41. Survey Paper - UCE15 MySQLMySQL implements asynchronous master-slave replication Masters Degree in Computer Engineering University of Minho
  42. 42. Survey Paper - UCE15 MySQL MySQL implements asynchronous master-slave replication Uses the primary-copy replication method, supporting two formatsof replication: Masters Degree in Computer Engineering University of Minho
  43. 43. Survey Paper - UCE15 MySQL MySQL implements asynchronous master-slave replication Uses the primary-copy replication method, supporting two formatsof replication: Statement-based Masters Degree in Computer Engineering University of Minho
  44. 44. Survey Paper - UCE15 MySQL MySQL implements asynchronous master-slave replication Uses the primary-copy replication method, supporting two formatsof replication: Statement-based Row-based Masters Degree in Computer Engineering University of Minho
  45. 45. Survey Paper - UCE15 MySQLReplication topologies: Masters Degree in Computer Engineering University of Minho
  46. 46. Survey Paper - UCE15 MySQL Replication topologies:Master and Multiple Slaves Masters Degree in Computer Engineering University of Minho
  47. 47. Survey Paper - UCE15 MySQL Replication topologies:Master and Multiple Slaves Ring Masters Degree in Computer Engineering University of Minho
  48. 48. Survey Paper - UCE15 MySQL Replication topologies:Master and Multiple Slaves Ring Tree Masters Degree in Computer Engineering University of Minho
  49. 49. Survey Paper - UCE15 ConclusionsMasters Degree in Computer Engineering University of Minho
  50. 50. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systems Masters Degree in Computer Engineering University of Minho
  51. 51. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occur Masters Degree in Computer Engineering University of Minho
  52. 52. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occurPrimary-copy approach introduces a single point of failure, but simplifies replica control Masters Degree in Computer Engineering University of Minho
  53. 53. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occurPrimary-copy approach introduces a single point of failure, but simplifies replica control Update-everywhere method speeds up data access but makes replica coordination morecomplex and expensive Masters Degree in Computer Engineering University of Minho
  54. 54. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occurPrimary-copy approach introduces a single point of failure, but simplifies replica control Update-everywhere method speeds up data access but makes replica coordination morecomplex and expensive1-copy serializability limits concurrency and thus scalability of the system Masters Degree in Computer Engineering University of Minho
  55. 55. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occurPrimary-copy approach introduces a single point of failure, but simplifies replica control Update-everywhere method speeds up data access but makes replica coordination morecomplex and expensive1-copy serializability limits concurrency and thus scalability of the system Concurrency control techniques and consistency criterion in lazy replication systems do notprovide data freshness Masters Degree in Computer Engineering University of Minho
  56. 56. Survey Paper - UCE15 ConclusionsEager protocols are not appropriate for large scale systemsLazy protocols have better performance, but inconsistencies among copies may occurPrimary-copy approach introduces a single point of failure, but simplifies replica control Update-everywhere method speeds up data access but makes replica coordination morecomplex and expensive1-copy serializability limits concurrency and thus scalability of the system Concurrency control techniques and consistency criterion in lazy replication systems do notprovide data freshness Some consistency techniques to improve data freshness emerged, but have a trade-offbetween consistency and performance Masters Degree in Computer Engineering University of Minho

×