Lecture--24
Lecture
Parallel and Distributed Databases
Contents
Introduction
Centralized and Client-Server Systems
Parallel
P ll l S
Systems
t
Distributed Systems
Applications
Scope of Research
Introduction
Centralized Systems
Run on a single computer system and do not interact with other
computer systems.
General-purpose computer system: one to a few CPUs and a number
of device controllers that are connected through a common bus that
provides access to shared memory.
p
y
Single-user system (e.g., personal computer or workstation): desk-top
unit, single user, usually has only one CPU and one or two hard
disks; the OS may support only one user.
Multi-user system: more disks, more memory, multiple CPUs, and a
multi-user OS. Serve a large number of users who are connected to
the system vie terminals. Often called server systems.
A Centralized Computer System
Client--Server Systems
Client
Server systems satisfy requests generated at m client systems, whose general
structure is shown below:
Client--Server Systems (Cont.)
Client
Database functionality can be divided into:
z
Back-end: manages access structures, query evaluation and
optimization concurrency control and recovery
optimization,
recovery.
Front-end: consists of tools such as forms, report-writers, and
graphical user interface facilities.
The interface between the front
front-end
end and the back
back-end
end is through SQL or
through an application program interface.
Client--Server Systems (Cont.)
Client
Advantages of replacing mainframes with networks of workstations or
personal computers connected to back-end server machines:
z
better functionality for the cost
flexibility in locating resources and expanding facilities
better user interfaces
easier
i maintenance
i t
Parallel Systems
Parallel database systems consist of multiple processors and multiple
disks connected by a fast interconnection network.
A coarse
coarse-grain
grain parallel machine consists of a small number of
powerful processors
A massively parallel or fine grain parallel machine utilizes
thousands of smaller processors.
p
Two main performance measures:
z
throughput --- the number of tasks that can be completed in a
given time interval
g
response time --- the amount of time it takes to complete a single
task from the time it is submitted
Speed--Up and ScaleSpeed
Scale-Up
Speedup: a fixed-sized problem executing on a small system is given
to a system which is N-times larger.
z
Measured by:
speedup = small system elapsed time
large system elapsed time
z
S
Speedup
d iis linear
li
if equation
ti equals
l N
N.
Scaleup: increase the size of both the problem and the system
z
N-times larger system used to perform N-times larger job
Measured by:
scaleup = small system small problem elapsed time
big system big problem elapsed time
z
Scale up is linear if equation equals 1.
Speedup
Speedup
Scaleup
Scaleup
p
Factors Limiting Speedup and Scaleup
Speedup and scaleup are often sublinear due to:
Startup costs: Cost of starting up multiple processes may dominate
computation time, if the degree of parallelism is high.
Interference: Processes accessing shared resources (e.g.,system
bus, disks, or locks) compete with each other, thus spending time
waiting
g on other p
processes, rather than p
performing
g useful work.
Skew: Increasing the degree of parallelism increases the variance in
service times of parallely executing tasks. Overall execution time
determined by slowest of parallely executing tasks.
Parallel Database Architectures
Shared memory -- processors share a common memory
Shared disk -- processors share a common disk
Shared nothing -- processors share neither a common memory nor
common disk
Hierarchical -- hybrid of the above architectures
Parallel Database Architectures
Shared Memory
Processors and disks have access to a common memory, typically via
a bus or through an interconnection network.
Extremely efficient communication between processors data in
shared memory can be accessed by any processor without having to
move it using software.
Downside architecture is not scalable beyond
y
32 or 64 processors
p
since the bus or the interconnection network becomes a bottleneck
Widely used for lower degrees of parallelism (4 to 8).
Shared Disk
All processors can directly access all disks via an interconnection
network, but the processors have private memories.
z
The memory bus is not a bottleneck
Architecture provides a degree of fault-tolerance if a processor
fails, the other processors can take over its tasks since the database
is resident on disks that are accessible from all processors.
Examples: IBM Sysplex and DEC clusters (now part of Compaq)
running Rdb (now Oracle Rdb) were early commercial users
Downside: bottleneck now occurs at interconnection to the disk
subsystem.
Shared-disk systems can scale to a somewhat larger number of
processors,, but communication between processors
p
p
is slower.
Shared Nothing
Node consists of a processor, memory, and one or more disks.
Processors at one node communicate with another processor at
another node using an interconnection network. A node functions as
the server for the data on the disk or disks the node owns.
Examples: Teradata, Tandem, Oracle-n CUBE
Data accessed from local disks ((and local memory
y accesses)) do not
pass through interconnection network, thereby minimizing the
interference of resource sharing.
Shared-nothing multiprocessors can be scaled up to thousands of
processors without interference.
Main drawback: cost of communication and non-local disk access;
sending data involves software interaction at both ends.
Hierarchical
Combines characteristics of shared-memory, shared-disk, and shared-
nothing architectures.
Top level is a shared
shared-nothing
nothing architecture nodes connected by an
interconnection network, and do not share disks or memory with each
other.
Each node of the system
y
could be a shared-memoryy system
y
with a
few processors.
Alternatively, each node could be a shared-disk system, and each of
the systems sharing a set of disks could be a shared-memory system.
Reduce the complexity of programming such systems by distributed
virtual-memory architectures
z
Also called non-uniform memory architecture (NUMA)
Distributed Systems
Data
D t spread
d over multiple
lti l machines
hi
((also
l referred
f
d tto as sites
it or
nodes).
Network interconnects the machines
Data shared by users on multiple machines
Distributed Databases
Homogeneous distributed databases
z
Same software/schema on all sites, data may be partitioned
among sites
z Goal: provide a view of a single database, hiding details of
distribution
Heterogeneous distributed databases
z Different software/schema on different sites
z Goal: integrate existing databases to provide useful functionality
Differentiate between local and global transactions
z A local transaction accesses data in the single site at which the
transaction was initiated.
z
A global transaction either accesses data in a site different from
th one att which
the
hi h the
th transaction
t
ti was initiated
i iti t d or accesses d
data
t iin
several different sites.
Trade--offs in Distributed Systems
Trade
Sharing data users at one site able to access the data residing at
some other sites.
Autonomy each site is able to retain a degree of control over data
stored locally.
Higher system availability through redundancy data can be
replicated
p
at remote sites, and system
y
can function even if a site fails.
Disadvantage: added complexity required to ensure proper
coordination among sites.
z
Software development
p
cost.
Greater potential for bugs.
Increased processing overhead.
Applications
The use of parallel database systems to deliver high performance has become
quite common. Although queries submitted to these database systems are
executed in parallel,
parallel the interaction between applications and current parallel
database systems is serial.
As the complexity of the applications and the amount of data they access
increases the need to parallelize applications also increases.
increases,
increases In this parallel
application environment, a serial interface to the database could become the
bottleneck in the performance of the application. Hence, parallel database
systems should support interfaces that allow the applications to interact with the
database system in parallel.
parallel
parallel database technology can contribute to make large and scalable
document management systems work.
Scope of research
Mobile, Service, P2P, grid and cloud computing for managing data and processes
Managing Heterogeneity and Autonomy in Distributed Systems
Semantic interoperability and integration (matching, mapping)
Linked
Linked Data
Data, Open Data
Data, Mobile Data
Data, Streaming Data
Data, Sensor Data
Data, M
Multimedia
ltimedia
and Multimodal Data
Metadata, Knowledge Bases, Ontologies
Web scale data management
Relational, Object-Oriented, XML, Graph, RDF, Event data management
