Parallel Databases


Parallel Databases
􀂄 Introduction
􀂄 I/O Parallelism
􀂄 Interquery Parallelism
􀂄 Intraquery Parallelism
􀂄 Intraoperation Parallelism
􀂄 Interoperation Parallelism
􀂄 Design of Parallel Systems

Introduction
􀂄 Parallel machines are becoming quite common and affordable
􀃌 Prices of microprocessors, memory and disks have dropped sharply
􀂄 Databases are growing increasingly large
􀃌 large volumes of transaction data are collected and stored for later
analysis.
􀃌 multimedia objects like images are increasingly stored in databases
􀂄 Large-scale parallel database systems increasingly used for:
􀃌 storing large volumes of data
􀃌 processing time-consuming decision-support queries
􀃌 providing high throughput for transaction processing

Parallelism in Databases
􀂄 Data can be partitioned across multiple disks for parallel I/O.
􀂄 Individual relational operations (e.g., sort, join, aggregation) can
be executed in parallel
􀃌 data can be partitioned and each processor can work independently
on its own partition.
􀂄 Queries are expressed in high level language (SQL, translated to
relational algebra)
􀃌 makes parallelization easier.
􀂄 Different queries can be run in parallel with each other.
Concurrency control takes care of conflicts.
􀂄 Thus, databases naturally lend themselves to parallelism.

I/O Parallelism
􀂄 Reduce the time required to retrieve relations from disk by partitioning
􀂄 the relations on multiple disks.
􀂄 Horizontal partitioning – tuples of a relation are divided among many
disks such that each tuple resides on one disk.
􀂄 Partitioning techniques (number of disks = n):
Round-robin:
Send the ith tuple inserted in the relation to disk i mod n.
Hash partitioning:
􀃌 Choose one or more attributes as the partitioning attributes.
􀃌 Choose hash function h with range 0…n - 1
􀃌 Let i denote result of hash function h applied tothe partitioning attribute
value of a tuple. Send tuple to disk i.

I/O Parallelism (Cont.)
􀂄 Partitioning techniques (cont.):
􀂄 Range partitioning:
􀃌 Choose an attribute as the partitioning attribute.
􀃌 A partitioning vector [vo, v1, ..., vn-2] is chosen.
􀃌 Let v be the partitioning attribute value of a tuple. Tuples such that vi
≤ vi+1 go to disk I + 1. Tuples with v < v0 go to disk 0 and tuples with
v ≥ vn-2 go to disk n-1.
E.g., with a partitioning vector [5,11], a tuple with partitioning attribute
value of 2 will go to disk 0, a tuple with value 8 will go to disk 1,
while a tuple with value 20 will go to disk2.

Comparison of Partitioning Techniques
􀂄 Evaluate how well partitioning techniques support the following
types of data access:
1.Scanning the entire relation.
2.Locating a tuple associatively – point queries.
􀃌 E.g., r.A = 25.
3.Locating all tuples such that the value of a given attribute lies
within a specified range – range queries.
􀃌 E.g., 10 ≤ r.A < 25.

Comparison of Partitioning Techniques (Cont.)
Round robin:
􀂄 Advantages
􀃌 Best suited for sequential scan of entire relation on each query.
􀃌 All disks have almost an equal number of tuples; retrieval work is
thus well balanced between disks.
􀂄 Range queries are difficult to process
􀃌 No clustering -- tuples are scattered across all disks

Comparison of Partitioning Techniques(Cont.)
Hash partitioning:
􀂄 Good for sequential access
􀃌 Assuming hash function is good, and partitioning attributes form a
key, tuples will be equally distributed between disks
􀃌 Retrieval work is then well balanced between disks.
􀂄 Good for point queries on partitioning attribute
􀃌 Can lookup single disk, leaving others available for answering other
queries.
􀃌 Index on partitioning attribute can be local to disk, making lookup
and update more efficient
􀂄 No clustering, so difficult to answer range queries

Comparison of Partitioning Techniques (Cont.)
Range partitioning:
􀂄 Provides data clustering by partitioning attribute value.
􀂄 Good for sequential access
􀂄 Good for point queries on partitioning attribute: only one disk
needs to be accessed.
􀂄 For range queries on partitioning attribute, one to a few disks
may need to be accessed
− Remaining disks are available for other queries.
− Good if result tuples are from one to a few blocks.
− If many blocks are to be fetched, they are still fetched from one
to a few disks, and potential parallelism in disk access is wasted
􀃌 Example of execution skew.

Partitioning a Relation across Disks
􀂄 If a relation contains only a few tuples which will fit into a single
disk block, then assign the relation to a single disk.
􀂄 Large relations are preferably partitioned across all the available
disks.
􀂄 If a relation consists of m disk blocks and there are n disks
available in the system, then the relation should be allocated
min(m,n) disks.









Comments

Post a Comment

Popular posts from this blog

Handling of Skew

Handling of Skew 􀂄 The distribution of tuples to disks may be skewed — that is, some disks have many tuples, while others may have fewer tuples. 􀂄 Types of skew: 􀃌 Attribute-value skew. 􀂾 Some values appear in the partitioning attributes of many tuples; all the tuples with the same value for the partitioning attribute end up in the same partition. 􀂾 Can occur with range-partitioning and hash-partitioning. 􀃌 Partition skew. 􀂾 With range-partitioning, badly chosen partition vector may assign too many tuples to some partitions and too few to others. 􀂾 Less likely with hash-partitioning if a good hash-function is chosen. Handling Skew in Range-Partitioning 􀂄 To create a balanced partitioning vector (assuming partitioning attribute forms a key of the relation): 􀃌 Sort the relation on the partitioning attribute. 􀃌 Construct the partition vector by scanning the relation in sorted order as follows. 􀂾 After every 1/nth of the relation has been read, the value...
Read more

Fragment-and-Replicate Join

Fragment-and-Replicate Join 􀂄 Partitioning not possible for some join conditions 􀃌 e.g., non-equijoin conditions, such as r.A > s.B. 􀂄 For joins were partitioning is not applicable, parallelization can be accomplished by fragment and replicate technique 􀃌 Depicted on next slide 􀂄 Special case – asymmetric fragment-and-replicate: 􀃌 One of the relations, say r, is partitioned; any partitioning technique can be used. 􀃌 The other relation, s, is replicated across all the processors. 􀃌 Processor Pi then locally computes the join of ri with all of s using any join technique. 􀂄 General case: reduces the sizes of the relations at each processor. 􀃌 r is partitioned into n partitions,r0, r1, ..., r n-1;s is partitioned into m partitions, s0, s1, ..., sm-1. 􀃌 Any partitioning technique may be used. 􀃌 There must be at least m * n processors. 􀃌 Label the processors as 􀃌 P0,0, P0,1, ..., P0,m-1, P1,0, ..., Pn-1m-1. 􀃌 Pi,j computes the join of ri with sj. In ...
Read more

USER INTERFACE DESIGN FOR ANNA UNIVERSITY SYLLABUS

Unit – I Introduction Objective To Understand the definition and importance of User Interface Design To know about the concept, advantages and applications of GUI and Web interface User Interface Design Definition • User interface design is a subset of a field of study called human-computer interaction (HCI). Human-computer interaction is the study, planning, and design of how people and computers work together so that a person’s needs are satisfied in the most effective way. • The user interface is the part of a computer and its software that people can see, hear, touch, talk to, or otherwise understand or direct. The user interface has essentially two components: input and output. • Input is how people communicate his needs to the system using keyboard or any pointing device and output is how the system returns processing result to user through screen or sound. • The best interface is one which has proper design with combination of effective input and output mechanisms. Importance of...
Read more