Distributed Systems

REKs adaptation of Prof. Claypools adaptation of Tanenbaums Distributed Systems Chapter 1

1

The Rise of Distributed Systems

Computer hardware prices are falling and power increasing. Network connectivity is increasing.

Everyone is connected with fat pipes.

It is easy to connect hardware together. Definition: a distributed system is

A collection of independent computers that appears to its users as a single coherent system.

Distributed Computing Systems

Forms of Transparency in a Distributed System

Transparency Access Location Migration Relocation Replication Concurrency Failure Persistence Description Hide differences in data representation and how a resource is accessed Hide where a resource is located Hide that a resource may move to another location

Hide that a resource may be moved to another location while in use

Hide that a resource may be shared by several competitive users Hide that a resource may be shared by several competitive users Hide the failure and recovery of a resource Hide whether a (software) resource is in memory or on disk
Distributed Computing Systems

Scalability Problems
Concept Centralized services Centralized data Centralized algorithms Example A single server for all users A single on-line telephone book Doing routing based on complete information

As distributed systems grow, centralized solutions are limited.

Distributed Computing Systems

Hiding Communication Latency

This is especially important for interactive applications

If possible, system can do asynchronous communication.

The system can hide latencies.
Distributed Computing Systems

Dividing the DNS name space into zones

1.5

Distributed Computing Systems

Hardware Concepts
Basic organizations and memories in distributed computer systems

1.6

Distributed Computing Systems

Hardware Considerations

General Classification:

Multiprocessor a single address space among the processors Multicomputer each machine has its own private memory.

OS can be developed for either type of environment.

Distributed Computing Systems

Multiprocessor Organizations

Uniform Memory Access [UMA]

Caching is vital for reasonable performance (e.g., caches on a shared memory multiprocessor). Want to maintain cache coherency

Write-through cache :: any changes to cache are written through to memory.

Distributed Computing Systems

Multiprocessors

1.7

A bus-based multiprocessor.

Distributed Computing Systems

10

Multiprocessors

A crossbar switch

An omega switching network

Distributed Computing Systems

11

Multiprocessor Organizations

Non-Uniform Memory Access [NUMA]

A hierarchy where CPUs have their own memory (not the same as a cache). Access costs to memory is non-uniform.

Distributed Computing Systems

12

Replication

Make a copy of information to increase availability and decrease centralized load.

Example: P2P networks (Gnutella +) distribute copies uniformly or in proportion to use. Example: CDNs (Akamai) Example: Caching is a replication decision made by client.

Issue: Consistency of replicated information

Example: Web Browser cache

Distributed Computing Systems

13

Software Concepts

DOS (Distributed Operating Systems) NOS (Network Operating Systems) Middleware

Description Main Goal

System DOS

Tightly-coupled operating system for multi- Hide and manage processors and homogeneous hardware multicomputers resources
Loosely-coupled operating system for heterogeneous multicomputers (LAN and WAN) Additional layer atop of NOS implementing general-purpose services
Distributed Computing Systems

NOS

Offer local services to remote clients Provide distribution transparency

14

Middleware

Uniprocessor Operating Systems

Separating applications from operating system code through a microkernel

Can extend to multiple computers

Distributed Computing Systems

15

Network Operating System

OSes can be different (Windows or Linux) Typical services: rlogin, rcp

Fairly primitive way to share files

Distributed Computing Systems

16

Network Operating System

Can have one computer provide files transparently for others (NFS)

(try a df on the WPI hosts to see. Similar to a mount network drive in Windows)
Distributed Computing Systems

17

Network Operating System

Different clients may mount the servers in different places Inconsistencies in view make NOSs harder, in general for users than DOSs. But easier to scale by adding computers
Distributed Computing Systems

18

Distributed Operating Systems

But no longer have shared memory

Provide message passing Can try to provide distributed shared memory

But tough to get acceptable performance

Distributed Computing Systems

19

Distributed System as Middleware

Distributed Computing Systems

20

Positioning Middleware

Network OSs are not transparent. Distributed OSs are not independent of computers.

Middleware can help.

Distributed Computing Systems

21

Middleware Models

View everything as a file - Plan 9. Less strict distributed file systems. Make all procedure calls appear to be local Remote Procedure Calls (RPC). Distributed objects (oo model). The Web distributed documents.

Distributed Computing Systems

22

Middleware and Openness

1.23

In an open middleware-based distributed system, the protocols used by each middleware layer should be the same, as well as the interfaces they offer to applications. If different, there will be compatibility issues If incomplete, then users will build their own or use lower-layer services (frowned upon)
Distributed Computing Systems

23

Comparison between Systems

Item Distributed OS Multiproc. Multicomp. Network OS Middleware -based OS

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

High
Yes N Messages Global, distributed Moderately Closed

Low
No N Files Per node Yes Open

High
No N Model specific Per node Varies Open

Distributed Computing Systems

24

Client-Server Model

Use TCP/IP for reliable network connection. This implies the client must establish a connection before sending the first request.

Distributed Computing Systems

25

Internet Search Engine

Distributed Computing Systems

26

Multitiered Architectures

Thin client (a) to Fat client (e)

(d) and (e) popular for NOS environments

Distributed Computing Systems

27

Multitiered Architectures: 3 tiers

Server may act as a client

Example would be transaction monitor across multiple databases

Distributed Computing Systems

28

Horizontal Distribution

Distribute servers across nodes

E.g., Web server farm for load balancing

Distribute clients in peer-to-peer systems.

Distributed Computing Systems

29