Professional Documents
Culture Documents
Arsitektur Server
Arsitektur Server
1
Arsitektur Server
Server
● A system which provides certain services to
another system, manages the use of resources
for the service, and responsible for the
computing
- File server
- Print server
- Network server
- Database server
● The “other system” can be in another computer
system, or in the same computer system
Server – Hardware Architecture
● There is no “real” limit for a server's hardware
specification
● Among all things, server is limited by microprocessor
performance
● Usually a hard working server employs
multiprocessors to serve its tasks
● Multiprocessor servers fall into one of two two major
classes:
a. Symmetrical Multiprocessor (Tightly coupled)
b. Loosely coupled
Server – HWA – Tightly Coupled
(SMP) Characteristics
● A computer system with many processors
● Each processor can access every resources of
the system (symmetrical)
● Access is managed by one operating system,
which must support multiprocessors, memory
sharing and threads (ex. Windows Server 2003)
● Creating a thread is running a different task for
the same process
Server – HWA – Tightly Coupled
Server – HWA – Tightly Coupled
Server – HWA – Tightly Coupled
Software Execution
Server – HWA – Loosely Coupled
Characteristics
● A collection of computer systems with
resources, interconnected
● Tasks are distributed among these collection of
hardware resources
Server – HWA – Loosely Coupled
Server – HWA – Loosely Coupled
Software Execution
Server – HWA – Performance
● Processor performance
● Memory capacity
● Throughput in connections within the system
(bus)
● Performance of storage peripherals
● Network troughput
Client
● A client system is a system which uses
service(s) of the server
● Relies on server for files, devices, processing
power
● “Clients” are applications that run on client
(user) systems, for example: an e-mail client.
Client – Server Computing
Server
Client – Server Computing
● Advantages:
- Easy to maintain
- Data consistency
- Resource efficiency
● Disadvantages:
- Single point of failure
- Possibility of data bottleneck
Generic Client–Server Architecture
Multiple Servers
Cluster Middleware
(Single System Image and Availability Infrastructure)