Overview

1
Introduction

• Chapter 1

2
Examples of Distributed Systems

• DNS
• Hierarchical distributed database
• WWW
• Origin servers and web caches
• Distributed database
• Cray T3E
• 2048 tightly coupled homogeneous processors
• Distributed/parallel computing
• Condor/RES
• Loosely coupled heterogeneous workstations
• Parallel/distributed computing

3
Other Distributed Systems

• Email
• Electronic banking
• Airline reservation system
• Peer-to-peer networks
• Etc., etc., etc.

4
Computer Revolution

• Processing power
• 50 years ago, 100M for 1 instr/sec
• Today, 1K for 107 instructions/sec
• Price/perform. improvement of 1012
• If cars had followed same path as computers
• a Rolls Royce would now cost 1 dollar and get a
  billion miles per gallon
• And it would explode once a year, killing
  everyone inside

5
Computer Revolution

• High speed networks
• 30 years ago, networks were unknown
• Today, Gigabit networks and the Internet
• Before networks, centralized systems
• Today, distributed systems
• Computers in many locations work as one

6
What is a Distributed System?

• According to your textbook
• A collection of independent computers that
  appears to its users as a single coherent system
  
• Two parts to definition
• Hardware ? machines are autonomous
• Software ? machines appear as one system
• Implies that communication hidden from user
• Implies that organization hidden from user

7
What is a Distributed System?

• According to dict.die.net
• A collection of (probably heterogeneous) automata
  whose distribution is transparent to the user so
  that the system appears as one local machine
• This is in contrast to a network, where the user
  is aware that there are several machines, and
  their location, storage replication, load
  balancing and functionality is not transparent
• Crucial point is transparency

8
How to Implement a Dist. System?

• A distributed system is a collection of
  independent computers
• that acts like a single system
• How to accomplish this?
• Middleware
• Make distributed system as transparent as possible

9
Role of Middleware

• Distributed system as middleware
• Middleware extends over multiple machines

10
Goals

• For a distributed system to be worthwhile authors
  believe it should
• Easily connect users to resources
• Hide fact that resources are distributed
• Be open
• Be scalable
• First 2 of these about transparency
• Transparent, open, scalable

11
Transparency

• Transparent system acts like one computer
• Various aspects of transparency listed above

12
Degree of Transparency

• Cannot hide physical limitations
• Time it takes to send packet
• May be a tradeoff between transparency and
  performance
• What to do if Web request times out?
• Keeping replicated data current

13
Openness

• Open standards-based
• Provides
• Interoperability
• Portability
• Ideally, flexible, i.e., extensible
• But many useful systems follow the American
  standard
• Do whatever you want

14
Scalability

• Scalability issues/limitations

15
Scalability

• Authors believe centralized is bad
• Centralized server is source of congestion,
  single point of failure
• Centralized data leads to congestion, lots of
  traffic
• Centralized algorithm must collect all info and
  process it (e.g., routing algs)
• Google? Napster?

16
Scalability

• Decentralized algorithms
• No machine has complete system state
• Decisions based on local info
• Failure of one machine does not kill entire
  algorithm
• No assumption of global clock
• Examples?

17
Geographic Scalability

• Big difference between LAN and WAN
• LANs have synchronous communication
• Client can block until server responds
• On LAN, global time may be possible (to within a
  few milliseconds)
• WAN unreliable, point-to-point
• WAN has different admin domains
• A security nightmare

18
Scaling Techniques

• Scaling problems due to limited capacity of
  networks and servers
• Three possible solutions
• Hide latencies ? do something useful while
  waiting (asynchronous comm.)
• Distribution ? DNS, for example
• Replication ? allows for load balancing
• Replication creates consistency issues

19
Scaling Techniques

• Server or client check form as its filled out?
• Having client do more, as in (b), may reduce
  latency (but may cause security problems)

20
Scaling Techniques

• DNS name space divided into zones
• Goto server in Z1 to find server Z2 and so on
• Like a binary search for correct server

21
Hardware Issues

• For our purposes, 2 kinds of machines
• Multiprocessor
• Different processors share same memory
• Multicomputer
• Each processor has its own memory
• Each of these could use either bus or switched
  architecture

22
Hardware Issues
multiprocessor
multicomputer
23
Multiprocessors

• A bus-based multiprocessor
• Cache coherence is an issue

24
Multiprocessors

• A crossbar switch
• Omega switching network

25
Homogeneous Multicomputer

• Grid

Hypercube
26
Software Concepts

• DOS ? Distributed Operating Systems
• NOS ? Network Operating Systems
• Middleware ? self-explanatory

27
Uniprocessor OSs

• Separate apps from OS code via microkernel

28
Multiprocessor OSs
monitor Counter private int count
0 public int value() return count void
incr () count count 1 void decr()
count count 1

• Hoe to protect count from concurrent access?

29
Multiprocessor OSs
monitor Counter private int count 0 int
blocked_procs 0 condition unblocked public
int value () return count void incr ()
if (blocked_procs 0) count
count 1 else signal
(unblocked)
void decr() if (count 0) blocked_procs
blocked_procs 1 wait (unblocked)
blocked_procs blocked_procs 1 else
count count 1

• Protect count from concurrent access
• Using blocking

30
Multicomputer OSs

• Multicomputer OS

31
Multicomputer OSs

• ???

32
Multicomputer OSs

• Huh?

33
Programming Issues

• Programming multicomputers much harder than
  multiprocessors
• Why?
• Message passing
• Buffering, blocking, reliable comm., etc.
• One option is to emulate shared memory on
  multicomputer
• Large virtual address space

34
Distributed Shared Memory

• Pages of address space distributed among 4
  machines
• After CPU 1 references pg 10
• If page 10 read only and replication used

35
Distributed Shared Memory

• False sharing of page between two processes
• Two independent processors share same page

36
Network OS

• Network OS
• Each processor has its own OS

37
Network OS

• Clients and server in a network OS
• Global shared file system

38
Distributed System

• Distributed OS not a distributed system by our
  definition
• Network OS not a distributed system by our
  definition
• What we need is middleware

39
Positioning Middleware

• A distributed system as middleware
• Individual node managed by local OS
• Middleware hides heterogeneity of underlying
  systems

40
Middleware and Openness

• Open middleware-based system
• Middleware layer should
• Use the same protocols
• Provide same interfaces to apps

41
Comparison of Systems

• Middleware rocks!

42
Middleware Services

• Main goal is access transparency
• Hides low level message passing
• Naming
• Like yellow pages or URL
• Persistence
• For example, a distributed file system
• Distributed transactions
• Read and writes are atomic
• Security

43
Client Server Model

• Read this section

44
Clients and Servers

• Interaction between client and server

45
Example Client and Server

• header.h
• Used by client
• And by server

46
Example Client and Server

• A sample server

47
Example Client and Server

• Client using server to copy a file

48
Processing Level

• Internet search engine as 3 layers

49
Multitiered Architectures

• Alternative client-server organizations

50
Multitiered Architectures

• A server acting as client

51
Modern Architectures

• Horizontal distribution of Web service

52
Summary

• Distributed system
• Autonomous computers that operate together as a
  single coherent system
• Potential advantages
• Can integrate homogeneous systems
• Scales well, if properly designed
• Potential disadvantages
• Complexity
• Degraded performance
• Poorer security

53
Summary

• Different types of dist systems
• Distributed OS
• For tightly coupled system
• Cant integrate different systems
• Network OS
• For heterogeneous system
• No single system view

54
Summary

• Middleware systems based on
• Remote procedure calls
• Distributed objects, files, documents
• Vertical organization
• Horizontal organization