Distributed Systems - Operating Systems - Study Guide | CMPSCI 377, Study notes of Operating Systems

Material Type: Notes; Professor: Berger; Class: Operating Systems; Subject: Computer Science; University: University of Massachusetts - Amherst; Term: Fall 2005;

Typology: Study notes

Pre 2010

Uploaded on 08/18/2009

koofers-user-8ky
koofers-user-8ky 🇺🇸

10 documents

1 / 4

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
CMPSCI 377 Operating Systems Fall 2005
Lecture 20: December 01
Lecturer: Emery Berger Scribe: Avery, Harris
Today:
Distributed Systems
Page Eviction
Network Topologies
Network Communication
Packets
20.1 Distributed Systems
20.1.1 General Information
A distributed system is a set of physically separate processors connected by one or more communication
links.
contains no shared clock or memory.
this systems are in broad use today.
20.1.2 Parallel vs. Distributed Systems
Tightly-coupled systems or ”parallel processing”
Systems share clocks, memory, etc.
Systems frequent communication
Loosely-coupled systems or ”distributed computing”
Each system/processor has its own memory and runs independent of the OS.
Systems have infrequent communication
20.1.3 Advantages of Distributed Systems
Resource Sharing
Resources do not need to be replicated because they are shared.
Expensive resources can be shared
Processors present the same environment to users.
20-1
pf3
pf4

Partial preview of the text

Download Distributed Systems - Operating Systems - Study Guide | CMPSCI 377 and more Study notes Operating Systems in PDF only on Docsity!

CMPSCI 377 Operating Systems Fall 2005

Lecture 20: December 01

Lecturer: Emery Berger Scribe: Avery, Harris

Today:

  • Distributed Systems
  • Page Eviction
  • Network Topologies
  • Network Communication
  • Packets

20.1 Distributed Systems

20.1.1 General Information

  • A distributed system is a set of physically separate processors connected by one or more communication links. - contains no shared clock or memory. - this systems are in broad use today.

20.1.2 Parallel vs. Distributed Systems

  • Tightly-coupled systems or ”parallel processing”
    • Systems share clocks, memory, etc.
    • Systems frequent communication
  • Loosely-coupled systems or ”distributed computing”
    • Each system/processor has its own memory and runs independent of the OS.
    • Systems have infrequent communication

20.1.3 Advantages of Distributed Systems

  • Resource Sharing
    • Resources do not need to be replicated because they are shared.
    • Expensive resources can be shared
    • Processors present the same environment to users.

20-2 Lecture 20: December 01

  • Computational Speedup
    • n processors, n times the computational power
    • Problems must be decomposable into subproblems
    • Coordination and communication is required between operating processes.
  • Reliability
    • Replication of resource produces fault tolerance. This cannot happen with only one machine.
    • If one node crashes the user can work on another.
    • Systems must be available for all but five minutes per year.
    • Must avoid single point of failure
  • Communication
    • Users, or processors, on different systems can communicate. Examples of this are mail servers and transaction processing systems.

20.1.4 Issues with Distributed Systems

  • Issues:
    1. Communication
    2. Transparency
    3. Security
    4. Reliability
    5. Performance
    6. Programming modules

20.2 Networks

20.2.1 General Information

  • Networks should provide efficient, correct, and robust message passing between two separate nodes.
  • Local area networks are connected in a singled building. This networks are extremely fast and reliable.
  • Wide area networks are connected across a large geographical area.

20.2.2 Network Topologies

  • Connection of the nodes impact the fault tolerance and cost of a network.
  • There are two basic topologies:
    1. Point-to-Point
    2. Bus

20-4 Lecture 20: December 01

20.2.4 Bus Network Topology

  • Bus nodes are connected to a common network. While this is very fast and reliable, it is not commonly utilized because of the poor scalability. - Linear Bus ∗ Single shared link. ∗ Nodes connect directly to each other via bus. ∗ Network is inexpensive and fault tolerant. - Ring Bus ∗ Single shared circular link. ∗ Same pros and cons as Linear Bus.

20.2.5 Principles of Network Communication

  • Data is broken down into packets. The packets are then sent through networks. These protocols are based on agreed communication standards.

20.2.6 Network Layers

  • Application: Data is used by applications.
  • Presentation: Data format conversion.
  • Session: Implement communication strategy.
  • Transport: Reliable end-to-end communication.
  • Network: Routing and congestion control.
  • Data Link Control: Reliable point-to-point communication over reliable channel.
  • Physical: Electrical signaling across wire.

20.2.7 TCP/IP Protocal Stack

  • TCP: Reliable protocol that ensures that packets are received in order.
  • UDP: Unreliable protocol that is fast, but does not ensure that packets are received in order.

20.2.8 Packets

  • Contain all information needed to recreate original message.
  • Packets may arrive out of order but can be put back in order through the sequence numbers of the packets.
  • Data segment contains header for higher protocol layers.