UNIFOR University of Fortaleza Brazil University of Kaiserslautern Germany  Angelo Brayner CoopIS - Trento, September Global Semantic Serializability: An Approach to Increase Concurrency in Multidatabase Systems Angelo Brayner Theo Härder

UNIFOR  Angelo BraynerCoopIS - Trento, September Contents Motivation Multidatabase System Model Global Semantic Serializability Concurrency Control Protocols Conclusions

UNIFOR  Angelo BraynerCoopIS - Trento, September Motivation (1) Integration of heterogeneous databases is a strategic requirement Integration of heterogeneous databases in a enterprise Integration of heterogeneous web databases  Web as a large collection of distributed autonomous and heterogeneous databases Integration of ubiquitous databases  mobile heterogeneous databases providing data everywhere

UNIFOR  Angelo BraynerCoopIS - Trento, September Motivation (2) Multidatabase technology Efficient solution for integrating a collection of autonomous and heterogeneous databases  Local databases Created independently without considering the possibility of being integrated in the future Operate autonomously Local autonomy is a key feature  Multidatabase Collection of local databases

UNIFOR  Angelo BraynerCoopIS - Trento, September Motivation (3) Multidatabase System (MDBS) Software component to manage a multidatabase Provides DBMS functionalities Multidatabase environment Global transactions  Submitted to the MDBS Access and update local database objects Local transactions  Submitted to local database systems

UNIFOR  Angelo BraynerCoopIS - Trento, September Motivation (4) Classical transaction-processing Model "Syntactic" serializability  Serialization order of all active transactions must be known For identifying correct execution of concurrent transactions Efficient criterion for synchronizing operations of short transactions

UNIFOR  Angelo BraynerCoopIS - Trento, September Motivation (5) Concurrency control problem in MDBSs Global transactions  Involve operations on multiple local databases Long-living transactions MDBS does not have any information about the execution (serialization) order of local transactions Classical transaction model is inefficient for solving the CC problem in MDBSs

UNIFOR  Angelo BraynerCoopIS - Trento, September Multidatabase System Model (1) LDBS n Local Transactions SUB jn SUB in Global Recovery Manager Global Log Global Transaction Manager MDBS GjGj GiGi Global Transactions Global Scheduler Interface Server 1 Log Interface Server n DB DBMS Local Transactions SUB i1 SUB j1 LDBS 1 DB DBMS

UNIFOR  Angelo BraynerCoopIS - Trento, September Multidatabase System Model (2) MDBS 1. A set LD ={LDBS 1, LDBS 2, …, LDBS n } of local database systems 2. A set L ={L 1, L 2, …, L n }  Each L K represents a set of local transactions executed at LDBS K 3. A set G ={G 1, G 2, …, G n } of global transactions

UNIFOR  Angelo BraynerCoopIS - Trento, September Multidatabase System Model (3) Local Schedule S K Models the execution of interleaved operations belonging to local and global transactions  Executed at LDBS K Global Schedule S G Models the execution of all local schedules

UNIFOR  Angelo BraynerCoopIS - Trento, September GS -Serializability Model (1) Assumptions An MDBS integrates a collection of pre- existing local databases (LDBs) A collection of disjoint sets of objects  Each set represents a single local database  Semantic Unit An update operation executed by a global transaction G on an object of a particular semantic unit does not depend on values of objects belonging to other semantic units previously read by G An update operation executed by a global transaction G on an object of a particular semantic unit does not depend on values of objects belonging to other semantic units previously read by G

UNIFOR  Angelo BraynerCoopIS - Trento, September GS -Serializability Model (2) Module-structured Transaction Operations are grouped into subsequences  Modules  Encompasses operations on objects of only one semantic unit Example  DB={A, B, C, D, E, F, G} A, B, C  SU LDBS1 ={A, B, C} D, E, F, G  SU LDBS2 ={D, E, F, G} 4 T 1 = r 1 (G) w 1 (E) w 1 (C) r 1 (B) 8 T 2 = r 2 (G) w 2 (C) w 2 (E) r 1 (B) Module

UNIFOR  Angelo BraynerCoopIS - Trento, September GS -Serializability Model (3) GS -Serial Global Schedule Local schedules are conflict serializable and Serial execution of modules belonging to global transactions Example CBAB G 1 =r 1 (G)w 1 (E)w 1 (C)r 1 (B); G 2 =r 2 (A)w 2 (B)w 2 (D)r 2 (E) ABCB S C = r 2 (A)w 2 (B)r 1 (G)w 1 (E)w 2 (D)r 2 (E)w 1 (C)r 1 (B)  S C is GS -Serial ì S C is not conflict serializable

UNIFOR  Angelo BraynerCoopIS - Trento, September GS -Serializability Model (4) GS -Serial Schedules preserve multidatabase consistency Correctness criterion for MDBSs GS -Serializable Schedule S Local schedules are conflict serializable and The execution order of global transactions in S is conflict equivalent to the execution of a GS -Serial schedule over the same set of transactions

UNIFOR  Angelo BraynerCoopIS - Trento, September GS -Serializability Model (5) Identifying GS -Serializable Schedule Since existing DBMSs yield conflict serializable schedules + The GTM has solely to verify the execution order of global transactions  A graph-based method The Semantic Serialization Graph (SSG)

UNIFOR  Angelo BraynerCoopIS - Trento, September Concurrency Control in MDBSs Concurrency Control Protocols Conservative  Based on a locking mechanism Aggressive  Management of an always acyclic graph Based on the SSG

UNIFOR  Angelo BraynerCoopIS - Trento, September Conclusions GS -Serializability Model Increases concurrency in MDBSs  More permissive than syntactic serializability Increases concurrency in mediator-based systems  Each web database can be seen as a semantic unit Can be applied to control concurrency in ubiquitous database  Mobile database can be defined as a semantic unit