DISTRIBUTED OPERATING SYSTEMS

Soham Ganguly University Institute Of Technology 2008-1054

DISTRIBUTED SYSTEM
´

´ ´ ´ ´

A distributed system is a collection of autonomous computers that appear to the users of the system as a single computer. « Multiple connected CPUs working together « A collection of independent computers that appears to its users as a single coherent system Runs on a cluster of machines with no shared memory Users get the feel of a single processor - virtual uniprocessor Transparency is the driving force. Requires « A single global IPC mechanism « A global protection mechanism « Identical process management and system calls at all nodes « Common file system at all nodes Examples: parallel machines, networked machines

´

WHY DISTRIBUTED SYSTEMS?
Resource sharing ´ Scalability ´ Reliability ´ Need for higher processing speed ´ Spatial distribution is inherent applications
´

in

many

HARDWARE CONCEPTS: MULTIPROCESSORS (1)
´

Multiprocessor dimensions
« «

Memory: could be shared or be private to each CPU Interconnect: could be shared (bus-based) or switched

´

A bus-based multiprocessor.

MULTIPROCESSORS (2)
(a) A crossbar switch A crossbar switch (also known as cross-point switch, crosspoint switch, or matrix switch) is a switch connecting multiple inputs to multiple outputs in a matrix manner. (b) An omega switching network An Omega network is a network configuration often used in parallel computing architectures. It is an indirect topology that relies on the perfect shuffle interconnection algorithm.

HOMOGENEOUS MULTICOMPUTER SYSTEMS
a)

Grid

b) Hypercube

DISTRIBUTED SYSTEMS MODELS
´

´

´

´

Minicomputer model (e.g., early networks) « Local processing but can fetch remote data (files, databases) « It consists of a few minicomputers each with multiple users. « Simple outgrowth of the Central Time-Sharing Systems. « Each user is locally logged-on to one machine and remotely logged-on to other machines. « (Logged-in Users / Available CPUs) < 1 Workstation model (e.g., Sprite) « Processing can also migrate « Each user has his personal workstation and nearly all work is done on the workstation. « Each user is locally logged-on to one machine and remotely logged-on to other machines. « It supports single, global file-system that provides location-independent data access. « (Logged-in Users / Available CPUs) ~ 1 Client-server Model (e.g., V system, world wide web) « User has local workstation « Powerful workstations serve as servers (file, print, DB servers) Processor pool model (e.g., Amoeba, Plan 9) « Terminals are Xterms or diskless terminals « When a user needs to perform computation, a processor is allocated from the processor pool to the user task. « (Logged-in Users / Available CPUs) > 1 « Pool of backend processors handle processing

TYPES OF DISTRIBUTED OSS

System DOS NOS Middleware

Description Tightly-coupled operating system for multiprocessors and homogeneous multicomputers Loosely-coupled operating system for heterogeneous multicomputers (LAN and WAN) Additional layer atop of NOS implementing generalpurpose services

Main Goal Hide and manage hardware resources Offer local services to remote clients Provide distribution transparency

In Multicomputer Operating System:‡ User is not aware of the multiple CPUs. ‡ Each machine runs a part of the Distributed Operating System. ‡ The system is fault-tolerant.

MULTICOMPUTER OPERATING SYSTEMS

NETWORK OPERATING SYSTEM
´It provides an environment where users are aware of the multiplicity of machines. ´Users can access remote resources by «logging into the remote machine OR «transferring data from the remote machine to their own machine ´Users should know where the required files and directories are and mount them. ´E.g. NFS from Sun Microsystems

NETWORK OPERATING SYSTEM
´

Employs a client-server model
« Minimal

OS kernel « Additional functionality as user processes

ADVANTAGES AND DISADVANTAGES
´

Advantages
« Communication

and resource sharing possible « Economics ² price-performance ratio « Reliability, scalability « Potential for incremental growth
´

Disadvantages
« Distribution-aware

PLs, OSs and applications « Network connectivity essential « Security and privacy

COMPARISON BETWEEN SYSTEMS
Distributed OS Item Multiproc. Degree of transparency Same OS on all nodes Number of copies of OS Basis for communication Resource management Scalability Openness Very High Yes 1 Shared memory Global, central No Closed Multicomp. High Yes N Messages Global, distributed Moderately Closed Low No N Files Per node Yes Open Network OS Middlewarebased OS High No N Model specific Per node Varies Open