Monday, December 3, 2012

Database Management System Chapter-17

 

 

Chapter 16: Concurrency Control

 

n Lock-Based Protocols

n Timestamp-Based Protocols

n Validation-Based Protocols

n Multiple Granularity

n Multiversion Schemes

n Deadlock Handling

n Insert and Delete Operations

n Concurrency in Index Structures

 

Lock-Based Protocols

 

n A lock is a mechanism to control concurrent access to a data item

n Data items can be locked in two modes :

    1exclusive (X) mode. Data item can be both read as well as  

         written. X-lock is requested using  lock-X instruction.

    2shared (S) mode. Data item can only be read. S-lock is         

         requested using  lock-S instruction.

n Lock requests are made to concurrency-control manager. Transaction can proceed only after request is granted.

 

n Lock-compatibility matrix

 

 

 

 

 

 

 

 


n A transaction may be granted a lock on an item if the requested lock is compatible with locks already held on the item by other transactions

n Any number of transactions can hold shared locks on an item, but if any transaction holds an exclusive on the item no other transaction may hold any lock on the item.

n If a lock cannot be granted, the requesting transaction is made to wait till all incompatible locks held by other transactions have been released.  The lock is then granted.

 

n  Example of a transaction performing locking:

                       T2: lock-S(A);

                             read (A);

                             unlock(A);

                             lock-S(B);

                             read (B);

                             unlock(B);

                             display(A+B)

n  Locking as above is not sufficient to guarantee serializability — if A and B get updated in-between the read of A and B, the displayed sum would be wrong.

n  locking protocol is a set of rules followed by all transactions while requesting and releasing locks. Locking protocols restrict the set of possible schedules.

 

Pitfalls of Lock-Based Protocols

 

n  Consider the partial schedule

 

 

 




 

 

 

 

 

 

 

 

 

n  Neither T3 nor T4 can make progress — executing  lock-S(B) causes T4 to wait for T3 to release its lock on B, while executing  lock-X(A) causes T3  to wait for T4 to release its lock on A.

n  Such a situation is called a deadlock.

é To handle a deadlock one of T3 or T4 must be rolled back
and its locks released.

n The potential for deadlock exists in most locking protocols. Deadlocks are a necessary evil.

n Starvation is also possible if concurrency control manager is badly designed. For example:

é A transaction may be waiting for an X-lock on an item, while a sequence of other transactions request and are granted an S-lock on the same item. 

é The same transaction is repeatedly rolled back due to deadlocks.

n Concurrency control manager can be designed to prevent starvation.

 

The Two-Phase Locking Protocol

 

n This is a protocol which ensures conflict-serializable schedules.

n Phase 1: Growing Phase

é transaction may obtain locks

é transaction may not release locks

n Phase 2: Shrinking Phase

é transaction may release locks

é transaction may not obtain locks

n The protocol assures serializability. It can be proved that the transactions can be serialized in the order of their lock points  (i.e. the point where a transaction acquired its final lock).

 

n Two-phase locking does not ensure freedom from deadlocks

n Cascading roll-back is possible under two-phase locking. To avoid this, follow a modified protocol called strict two-phase locking. Here a transaction must hold all its exclusive locks till it commits/aborts.

n Rigorous two-phase locking is even stricter: here all locks are held till commit/abort. In this protocol transactions can be serialized in the order in which they commit.

 

n There can be conflict serializable schedules that cannot be obtained if two-phase locking is used. 

n However, in the absence of extra information (e.g., ordering of  access to data), two-phase locking is needed for conflict serializability in the following sense:

    Given a transaction Ti that does not follow two-phase locking, we can find a transaction Tj that uses two-phase locking, and a schedule for Ti and Tj that is not conflict serializable.

 

 

 

Lock Conversions

 

n Two-phase locking with lock conversions:

     –   First Phase:       

é can acquire a lock-S on item

é can acquire a lock-X on item

é can convert a lock-S to a lock-X (upgrade)

     –   Second Phase:

é can release a lock-S

é can release a lock-X

é can convert a lock-X to a lock-S  (downgrade)

n This protocol assures serializability. But still relies on the programmer to insert the various  locking instructions.

 

Automatic Acquisition of Locks

 

n A transaction Ti issues the standard read/write instruction, without explicit locking calls.

n The operation read(D) is processed as:

                      if Ti has a lock on D

                         then

                                read(D)

                         else

                                begin

                                   if necessary wait until no other 

                                       transaction has a lock-X on D

                                   grant Ti a  lock-S on D;

                                   read(D)

                                end

 

 

n  write(D) is processed as:

     if Ti has a  lock-X on D

        then

          write(D)

       else

         begin

            if necessary wait until no other trans. has any lock on D,

            if Ti has a lock-S on D

                 then

                    upgrade lock on D  to lock-X

                else

                    grant Ti a lock-X on D

                write(D)

         end;

n  All locks are released after commit or abort

 

Implementation of Locking

 

n A Lock manager can be implemented as a separate process to which transactions send lock and unlock requests

n The lock manager replies to a lock request by sending a lock grant messages (or a message asking the transaction to roll back, in case of  a deadlock)

n The requesting transaction waits until its request is answered

n The lock manager maintains a datastructure called a lock table to record granted locks and pending requests

n The lock table is usually implemented as an in-memory hash table indexed on the name of the data item being locked

 

Lock Table

 

n  Black rectangles indicate granted locks, white ones indicate waiting requests

n  Lock table also records the type of lock granted or requested

n  New request is added to the end of the queue of requests for the data item, and granted if it is compatible with all earlier locks

n  Unlock requests result in the request being deleted, and later requests are checked to see if they can now be granted

n  If transaction aborts, all waiting or granted requests of the transaction are deleted

é lock manager may keep a list of locks held by each transaction, to implement this efficiently

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Graph-Based Protocols

 

n Graph-based protocols are an alternative to two-phase locking

n Impose a partial ordering ® on the set D = {d1, d2 ,..., dh} of all data items.

é If di ® dj  then any transaction accessing both di and dj must access di before accessing dj.

é Implies that the set D may now be viewed as a directed acyclic graph, called a database graph.

n The tree-protocol is a simple kind of graph protocol.

 

 

 

 

 

Tree Protocol

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


n Only exclusive locks are allowed.

n The first lock by Ti may be on any data item. Subsequently, a data Q can be locked by Ti only if the parent of Q is currently locked by Ti.

n Data items may be unlocked at any time.

 

n The tree protocol ensures conflict serializability as well as freedom from deadlock.

n Unlocking may occur earlier in the tree-locking protocol than in the two-phase locking protocol.

é shorter waiting times, and increase in concurrency

é protocol is deadlock-free, no rollbacks are required

é the abort of a transaction can still lead to cascading rollbacks.

     (this correction  has to be made in the book also.)

n However,  in the tree-locking protocol, a transaction may have to lock data items that it does not access.

é  increased locking overhead, and additional waiting time

é potential decrease in concurrency

n Schedules not possible under two-phase locking are possible under tree protocol, and vice versa.

 

Timestamp-Based Protocols

 

n Each transaction is issued a timestamp when it enters the system. If an old transaction Ti has time-stamp TS(Ti), a new transaction Tj is assigned time-stamp TS(Tj) such that TS(Ti) <TS(Tj).

n The protocol manages concurrent execution such that the time-stamps determine the serializability order.

n In order to assure such behavior, the protocol maintains for each data Q two timestamp values:

é W-timestamp(Q) is the largest time-stamp of any transaction that executed write(Q) successfully.

é R-timestamp(Q) is the largest time-stamp of any transaction that executed read(Q) successfully.

 

n The timestamp ordering protocol ensures that any conflicting  read and write operations are executed in timestamp order.

n Suppose a transaction Ti issues a read(Q)

  1.  If TS(Ti) £ W-timestamp(Q), then Ti needs to read a value of Q      

       that was already overwritten. Hence, the read operation is       

       rejected, and Ti  is rolled back.

  2.  If TS(Ti)³ W-timestamp(Q), then the read operation is

       executed, and R-timestamp(Q) is set to the maximum of R-

       timestamp(Q) and TS(Ti).

 

n Suppose that transaction Ti issues write(Q).

n If TS(Ti) < R-timestamp(Q), then the value of Q that Ti is producing was needed previously, and the system assumed that that value would never be produced. Hence, the write operation is rejected, and Ti is rolled back.

n If TS(Ti) < W-timestamp(Q), then Ti is attempting to write an obsolete value of Q. Hence, this write operation is rejected, and Ti is rolled back.

n Otherwise, the  write operation is executed, and W-timestamp(Q) is set to TS(Ti).

 

Example Use of the Protocol

 

A partial schedule for several data items for transactions with

timestamps 1, 2, 3, 4, 5

                  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Correctness of Timestamp-Ordering Protocol

n The timestamp-ordering protocol guarantees serializability since all the arcs in the precedence graph are of the form:

   

 

 

 


         

 

 

 

 

 

 

 

    Thus, there will be no cycles in the precedence graph

n Timestamp protocol ensures freedom from deadlock as no transaction ever waits. 

n But the schedule may not be cascade-free, and may  not even be recoverable.

 

Recoverability and Cascade Freedom

 

n Problem with timestamp-ordering protocol:

é Suppose Ti aborts, but Tj has read a data item written by  Ti

é Then Tj must abort; if Tj had been allowed to commit earlier, the schedule is not recoverable.

é Further, any transaction that has read a data item written by Tj must abort

é This can lead to cascading rollback --- that is, a chain of rollbacks

n  Solution:

é A transaction is structured such that its writes are all performed at the end of its processing

é All writes of a transaction form an atomic action; no transaction may execute while a transaction is being written

é A transaction that aborts is restarted with a new timestamp

 

Thomas’ Write Rule

 

n Modified version of the timestamp-ordering protocol in which obsolete  write operations may be ignored under certain circumstances.

n When Ti attempts to write data item Q, if TS(Ti) < W-timestamp(Q), then Ti is attempting to write an obsolete value of {Q}. Hence, rather than rolling back Ti as the timestamp ordering protocol would have done, this {write} operation can be ignored.

n Otherwise this protocol is the same as the timestamp ordering protocol.

n Thomas' Write Rule allows greater potential concurrency. Unlike previous protocols, it allows some view-serializable schedules that are not conflict-serializable.

Validation-Based Protocol

 

n Execution of transaction Ti is done in three phases.

  1.  Read and execution phase: Transaction Ti writes only to        

       temporary local variables

  2.  Validation phase: Transaction Ti performs a ``validation test''

        to determine if local variables can be written without violating        

        serializability.

  3.  Write phase: If Ti is validated, the updates are applied to the

      database; otherwise, Ti is rolled back.

n The three phases of concurrently executing transactions can be    interleaved, but each transaction must go through the three phases in that order.

n Also called as optimistic concurrency control since transaction executes fully in the hope that all will go well during validation

 

n Each transaction Ti has 3 timestamps

-      Start(Ti) : the time when Ti started its execution

-      Validation(Ti): the time when Ti entered its validation phase

-       Finish(Ti) : the time when Ti finished its write phase

n Serializability order is determined by timestamp given at validation time,  to increase concurrency. Thus TS(Ti) is given the value of Validation(Ti).

n This protocol is useful and gives greater degree of concurrency if probability of conflicts is low. That is because the serializability order is not pre-decided and relatively less transactions will have to be rolled back.

 

Validation Test for Transaction Tj

 

n If for all Ti with TS (Ti) < TS (Tj) either one of the following condition holds:

é finish(Ti) < start(Tj)

é start(Tj) < finish(Ti) < validation(Tj) and the set of data items written by Ti does not intersect with the set of data items read by Tj

     then validation succeeds and Tj can be committed.  Otherwise, validation fails and Tj is aborted.

n Justification:  Either first condition is satisfied, and there is no overlapped execution, or second condition is satisfied and

  1.  the writes of Tj do not affect reads of Ti since they occur after Ti

       has finished its reads.

  2.  the writes of Ti do not affect reads of Tj since Tj does not read 

      any item written by Ti.

 

Schedule Produced by Validation

 

n Example of schedule produced using validation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Multiple Granularity

 

n Allow  data items to be of various sizes and define a hierarchy of data granularities, where the small granularities are nested within larger ones

n Can be represented graphically as a tree (but don't confuse with tree-locking protocol)

n When a transaction locks a node in the tree explicitly, it implicitly locks all the node's descendents in the same mode.

n Granularity of locking (level in tree where locking is done):

é fine granularity (lower in tree): high concurrency, high locking overhead

é coarse granularity  (higher in tree): low locking overhead, low concurrency

 

 

 

 

Example of Granularity Hierarchy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


   The highest level in the example hierarchy is the entire database.

   The levels below are of type area, file and record in that order.

 

Intention Lock Modes

 

n In addition to S and X lock modes, there are three additional lock modes with multiple granularity:

é intention-shared (IS): indicates explicit locking at a lower level of the tree but only with shared locks.

é intention-exclusive (IX): indicates explicit locking at a lower level with exclusive or shared locks

é shared and intention-exclusive (SIX): the subtree rooted by that node is locked explicitly in shared mode and explicit locking is being done at a lower level with exclusive-mode locks.

n intention locks allow a higher level node to be locked in S or X mode without having to check all descendent nodes.

 

 

 

 

Compatibility Matrix with
Intention Lock Modes

 

n The compatibility matrix for all lock modes is:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Multiple Granularity Locking Scheme

 

n  Transaction Ti can lock a node Q, using the following rules:

  1. The lock compatibility matrix must be observed.

  2. The root of the tree must be locked first, and may be locked in

       any mode.

  3. A node Q can be locked by Ti in S or IS mode only if the parent

      of Q is currently locked by Ti in either IX or IS                         

      mode.

  4. A node Q can be locked by Ti in X, SIX, or IX mode only if the

      parent of Q is currently locked by Ti in either IX   

      or SIX mode.

  5. Ti can lock a node only if it has not previously unlocked any node

      (that is, Ti is two-phase).

  6. Ti can unlock a node Q only if none of the children of Q are

      currently locked by Ti.

n  Observe that locks are acquired in root-to-leaf order,
whereas they are released in leaf-to-root order.

 

Multiversion Schemes

 

n Multiversion schemes keep old versions of data item to increase concurrency.

é Multiversion Timestamp Ordering

é Multiversion Two-Phase Locking

n Each successful write results in the creation of a new version of the data item written.

n Use timestamps to label versions.

n When a read(Q) operation is issued, select an appropriate version of Q based on the timestamp of the transaction, and return the value of the selected version. 

n reads never have to wait as an appropriate version is returned immediately.

 

Multiversion Timestamp Ordering

 

n Each data item Q has a sequence of versions <Q1, Q2,...., Qm>. Each version Qk contains three data fields:

é Content -- the value of version Qk.

é W-timestamp(Qk) -- timestamp of the transaction that created (wrote) version Qk

é R-timestamp(Qk) -- largest timestamp of a transaction that successfully read version Qk

n when a transaction Ti creates a new version Qk of Q, Qk's W-timestamp and R-timestamp are initialized to TS(Ti).

n R-timestamp of Qk is updated whenever a transaction Tj reads Qk, and TS(Tj) > R-timestamp(Qk).

 

n The multiversion timestamp scheme presented next ensures serializability.

n Suppose that transaction Ti issues a read(Q) or write(Q) operation.  Let Qk denote the version of Q whose write timestamp is the largest write timestamp less than or equal to TS(Ti).

  1.  If transaction Ti issues a read(Q), then the value returned is the      

       content of version Qk.

  2.  If transaction Ti issues a  write(Q), and if TS(Ti) < R-

       timestamp(Qk), then transaction Ti is rolled

       back. Otherwise,  if TS(Ti) = W-timestamp(Qk), the contents of Qk

       are overwritten, otherwise a new version of Q is created.

n Reads always succeed; a write by Ti is rejected if some other transaction Tj that (in the serialization order defined by the timestamp values) should read Ti's write, has already read a version created by a transaction older than Ti.

 

 

Multiversion Two-Phase Locking

 

n Differentiates between read-only transactions and update transactions

n Update transactions acquire read and write locks, and hold all locks up to the end of the transaction. That is, update transactions follow rigorous two-phase locking.

é Each successful write results in the creation of a new version of the data item written.

é each version of a data item has a single timestamp whose value is obtained from a counter ts-counter that is incremented during commit processing.

n Read-only transactions are assigned a timestamp by reading the current value of  ts-counter before they start execution; they follow the multiversion timestamp-ordering protocol for performing reads.

 

n When an update transaction wants to read a data item, it obtains a shared lock on it, and reads the latest version.

n When it wants to write an item, it obtains X lock on; it then creates a new version of the item and sets this version's timestamp to ¥.

n When update transaction Ti completes, commit processing occurs:

é Ti sets timestamp on the versions it has created to  ts-counter + 1

é Ti increments  ts-counter by 1

n Read-only transactions that start after Ti increments ts-counter will see the values updated by Ti.

n Read-only transactions that start before Ti increments the
ts-counter will see the value before the updates by Ti.

n Only serializable schedules are produced.

 

 

Deadlock Handling

 

n Consider the following two transactions:

             T1:     write (X)               T2:    write(Y)

                      write(Y)                         write(X)

n Schedule with deadlock

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Deadlock Handling

 

n System is deadlocked if there is a set of transactions such that every transaction in the set is waiting for another transaction in the set.

n Deadlock prevention protocols ensure that the system will never enter into a deadlock state. Some prevention strategies :

é Require that each transaction locks all its data items before it begins execution (predeclaration).

é Impose partial ordering of all data items and require that a transaction can lock data items only in the order specified by the partial order (graph-based protocol).

More Deadlock Prevention Strategies

n Following schemes use transaction timestamps for the sake of deadlock prevention alone.

n wait-die scheme — non-preemptive

é older transaction may wait for younger one to release data item. Younger transactions never wait for older ones; they are rolled back instead.

é a transaction may die several times before acquiring needed data item

n wound-wait scheme — preemptive

é older transaction wounds (forces rollback) of younger transaction instead of waiting for it. Younger transactions may wait for older ones.

é may be fewer rollbacks than wait-die scheme.

n Both in wait-die and in wound-wait schemes, a rolled back transactions is restarted with its original timestamp. Older transactions thus have precedence over newer ones, and starvation is hence avoided.

n Timeout-Based Schemes :

é a transaction waits for a lock only for a specified amount of time. After that, the wait times out and the transaction is rolled back.

é thus deadlocks are not possible

é simple to implement; but starvation is possible. Also difficult to determine good value of the timeout interval.

 

Deadlock Detection

 

n Deadlocks can be described as a wait-for graph, which consists of a pair G = (V,E),

é V is a set of vertices (all the transactions in the system)

é E is a set of edges; each element is an ordered pair Ti ®Tj

n If Ti ®  Tj is in E, then there is a directed edge from Ti to Tj, implying that Ti is waiting for Tj to release a data item.

n When Ti requests a data item currently being held by Tj, then the edge Ti  Tj is inserted in the wait-for graph. This edge is removed only when Tj is no longer holding a data item needed by Ti.

n The system is in a deadlock state if and only if the wait-for graph has a cycle.  Must invoke a deadlock-detection algorithm periodically to look for cycles.

 

 

 

 

 

 

 

 

 

 

 

 

 

 


                                                                   

 

 

 

 

 

 

 

Wait-for graph without a cycle                            Wait-for graph with a cycle

 

 

 

 

 

Deadlock Recovery

 

n When deadlock is  detected :

é Some transaction will have to rolled back (made a victim) to break deadlock.  Select that transaction as victim that will incur minimum cost.

é Rollback -- determine how far to roll back transaction

Ø Total rollback: Abort the transaction and then restart it.

Ø More effective to roll back transaction only as far as necessary to break deadlock.

é Starvation happens if same transaction is always chosen as victim. Include the number of rollbacks in the cost factor to avoid starvation

 

Insert and Delete Operations

 

n If two-phase locking is used :

é A  delete operation may be performed only if the transaction deleting the tuple has an exclusive lock on the tuple to be deleted.

é A transaction that inserts a new tuple into the database is given an X-mode lock on the tuple

n Insertions and deletions can lead to the phantom phenomenon.

é A transaction that scans a relation (e.g., find all accounts in Perryridge) and a transaction that inserts a tuple in the relation (e.g., insert a new account at Perryridge) may conflict in spite of not accessing any tuple in common.

é If only tuple locks are used, non-serializable schedules can result: the scan transaction may not see the new account, yet may be serialized before the insert transaction.

 

n The transaction scanning the relation is reading  information that indicates what tuples the relation contains, while a transaction inserting a tuple updates the same information.

é  The information should be locked.

n One solution:

é Associate a data item with the relation, to represent the information about what tuples the relation contains.

é Transactions scanning the relation acquire a shared lock in the data item,

é Transactions inserting or deleting a tuple acquire an exclusive lock on the data item. (Note: locks on the data item do not conflict with locks on individual tuples.)

n Above protocol provides very low concurrency for insertions/deletions.

n Index locking protocols provide higher concurrency while
preventing the phantom phenomenon, by requiring locks
on certain index buckets.

 

Index Locking Protocol

 

n Every relation must have at least one index. Access to a relation must be made only through one of the indices on the relation.

n A transaction Ti that performs a lookup must lock all the index buckets that it accesses, in S-mode.

n A transaction Ti may not insert a tuple ti into a relation r  without updating all indices to r.

n Ti must perform a lookup on every index to find all index buckets that could have possibly contained a pointer to tuple ti, had it existed already, and obtain locks in X-mode on all these index buckets. Ti must also obtain locks in X-mode on all index buckets that it modifies.

n The rules of the two-phase locking protocol must be observed.

 

Weak Levels of Consistency

 

n Degree-two consistency: differs from two-phase locking in that S-locks may be released at any time, and locks may be acquired at any time

é X-locks must be held till end of transaction

é Serializability is not guaranteed, programmer must ensure that no erroneous database state will occur]

n Cursor stability:

é For reads, each tuple is locked, read, and lock is immediately released

é X-locks are held till end of transaction

é Special case of degree-two consistency

 

Weak Levels of Consistency in SQL

 

n SQL allows non-serializable executions

é Serializable: is the default

é Repeatable read: allows only committed records to be read, and repeating a read should return the same value (so read locks should be retained)

Ø However, the phantom phenomenon need not be prevented

  T1 may see some records inserted by T2, but may not see others inserted by T2

é Read committed:  same as degree two consistency, but most systems implement it as cursor-stability

é Read uncommitted: allows even uncommitted data to be read

 

Concurrency in Index Structures

 

n Indices are unlike other database items in that their only job is to help in accessing data.

n Index-structures are typically accessed very often, much more than other database items.

n Treating index-structures like other database items leads to low concurrency.   Two-phase locking on an index may result in transactions executing practically one-at-a-time.

n It is acceptable to have nonserializable concurrent access to an index as long as the accuracy of the index is maintained.

n In particular, the exact values read in an internal node of a
B+-tree are irrelevant so long as we land up in the correct leaf node.

n There are index concurrency protocols where locks on internal nodes are released early, and not in a two-phase fashion.

 

n Example of index concurrency protocol:

n Use crabbing instead of two-phase locking on the nodes of the B+-tree, as follows.  During search/insertion/deletion:

é First lock the root node in shared mode.

é After locking all required children of a node in shared mode, release the lock on the node.

é During insertion/deletion, upgrade leaf node locks to exclusive mode.

é When splitting or coalescing requires changes to a parent, lock the parent in exclusive mode.

n Above protocol can cause excessive deadlocks. Better protocols are available; see Section 16.9 for one such protocol, the B-link tree protocol

 

 

 

 

 

End of Chapter

 

 

 

0 comments:

Post a Comment