CO526 Distributed Systems and Networks

Card Set Information

CO526 Distributed Systems and Networks
2013-05-10 16:07:16
Distributed Systems Networks

Revision cards for CO526 Distributed Systems and Networks
Show Answers:

  1. State the name of each layer in the ISO seven-layer model
    • Layer 7 - Application
    • Layer 6 - Presentation
    • Layer 5 - Session
    • Layer 4 - Transport
    • Layer 3 - Network
    • Layer 2 - Data Link
    • Layer 1 - Physical
  2. Give some disadvantages of standardizing
    • Standardizing when research is incomplete, thus leading to poor standards.
    • Standards can be overly complex and thus restrictive on new people entering the market.
    • Standards may be incomplete.
    • Standards may become irrelevant as technology advances.
  3. State the purpose of the Physical layer in ISO seven-layer network model.
    Bit transmission across some sort of transmission meidum
  4. State the purpose of the Data Link layer in ISO seven-layer network model.
    Blocking bits together to control flow and checking that they arrived correctly
  5. State the purpose of the Network layer in ISO seven-layer network model.
    Routing blocks of bits through a network
  6. State the purpose of the Transport layer in ISO seven-layer network model.
    Receives the blocks at the destination (and passes them to the right application)
  7. State the purpose of the Session layer in ISO seven-layer network model.
    Synchronization and Dialogue control. Managing sessions between applications.
  8. State the purpose of the Presentation layer in ISO seven-layer network model.
    Syntax and semantics of the information. (Encrypting / Decrypting, Data representation)
  9. State the purpose of the Application layer in ISO seven-layer network model.
    Common applications
  10. Name the internet layers
    • Application Layer
    • Transport Layer
    • Internet Protocol Layer
    • Network Interface Layer
    • Physical Net Layer
  11. Which layers of the Intenet model correspond with that of the OSI seven-layer model.
    • Internet Application - OSI Application
    • Internet Application - OSI Presentation
    • Internet Application - OSI Session
    • Internet Transport - OSI Transport
    • Internet Internet Protocol - OSI Network
    • Internet Network Interface - OSI Data Link
    • Internet Physical - OSI Physical
  12. Why is layering (abstraction) a good idea?
    • Focus on the individual processes
    • Debugging and maintenance is easier
  13. What is a protocol?
    "A set of rules and regulations agreed by two parties in order to govern communication."
  14. What is framing and why is it used?
    It is where a raw bit stream is broken up into groups of bits (called frames). This is to enable error and flow control.
  15. Data Link Protocol: What would be an unrestricted protocol?
    A protocol whereby the sender could send frames as fast as it wanted. (Unrestricted send speed)
  16. Data Link Protocol: What would be a restricted protocol?
    • (Stop-and-wait protocol)
    • Sender sends a frame and then waits for an acknowledgement (ACK)
    • Receiver waits for a frame and then sends an acknowledgement (ACK) when it has received one.
  17. Data Link Protocol: What is put in place to help tackle a noisy channel?
    • Sender times out if it doesn't receive an acknowledgement (ACK). It will then re-send the same frame.
    • We use sequence numbers so that the receiver can check that the expected frame has been received. The receiver then sends an acknowledgement (ACK) back to the sender.
  18. The speed at which bits are transmitted depends on the ..(a).. . The distance bits can be transmitted depends on the ..(a).. AND the ..(b).. .
    • (a) Transmission Medium
    • (b) Protocol
  19. Give three positions where you can  interconnect protocols
    • At the Physical layer - Amplify the signals
    • At the Data Link layer - Regenerate the bits, change protocol
    • At the Network layer - Choose different routes across alternative networks.
  20. What are Repeaters?
    Repeaters operate at the Physical layer and amplify bits. (Forward them on)
  21. What does a multiport repeater do?
    • Extend LAN to several segments from a single location.
    • Can form the centre of a star network of hosts (Hubs)
  22. What are Switches?
    • Similar to selective repeaters but form a star network like Hubs.
    • Frames are sent only on the link to the destination
    • Half way between the Physical layer and the Data Link layer.
  23. What are Bridges?
    • They operate at Data Link layer
    • May connect similar or dissimilar LANs
    • Unmanaged bridges filter packets
    • Managed bridges also can take traffic measurements and be accessed remotely to examine their statistics.
  24. What are Routers?
    • These form the basis of internets.
    • They operate on the network layer.
    • Joins networks together.
  25. Describe a Connectionless Service
    • Each datagram (packet) is routed individually (each one carrying the full address of its destination)
    • Datagrams (packets) can arrive out of sequence.
    • Unreliable delivery of data.
  26. Describe a Connection Oriented Service
    • Path established before data is sent
    • Packets always arrive in order
    • Packets do not have to carry the full address (only the virtual circuit number is needed)
  27. Define Adaptive routing
    Routing decisions are modified during network operation
  28. How does non-adaptive routing work?
    • Each router maintains a table with a row for every destination
    • For each row there are several choice of route with a relative weight for each choice
    • Random numbers are used in conjunction with weights to determine route
  29. What is centralised routing?
    • Routers maintain tables as they would in non-adaptive routing.
    • Periodically routers will send reports to routing control centres (RCC)
    • RCC calculates and distributes new tables based on status reports
  30. What is isolated routing?
    Each router looks after its own routing without communicating with others.
  31. In isolated routing, what is 'Hot Potato' communicating?
    Place outgoing packet on the shortest queue
  32. In isolated routing, what is 'Flooding'?
    • Send incoming packets on all outgoing lines.
    • It will use a hop counter to prevent continuous circulation of packets
  33. How does link state routing work?
    • 1st - Find out who your neighbours are (Send "hello" packet on each line)
    • 2nd - Measure the delay to each neighbour (Use "Echo" packet)
    • 3rd - Send this information to all other routers (Use modified flooding)
    • 4th - Work out the shortest path to every router
    • 5th - Use the shortest path for routing
  34. What does DNS stand for?
    Domain Name Service
  35. What does FTP stand for?
    File Transfer Protocol
  36. What does SMTP stand for?
    Simple Main Transfer Protocol
  37. What does SNMP stand for?
    Simple Network Management Protocol
  38. What does HTTP stand for?
    Hypertext Transport Protocol
  39. What is the job of the DNS?
    • Translates names into addresses. Returns an IP address to the user.
    • Offers the use of host aliasing (canonical naming)
    • Stores the different IP's for the same site (that have been stored on different servers). Distributes load
  40. Why is there not just one DNS server? (DNS servers are spread globally)
    • Single point of failure
    • High traffic volumes
    • Users may be far away
    • It would contain an extremely large database
  41. What is the frame format for an Ethernet II Frame?
    • Dest. Address:     6
    • Source Address:   6
    • Ether-Type:         2
    • Data:                   46-1500
    • FCS:                    4

    FCS = Frame Check Sequence
  42. Which internet layers do these protocols sit?

    Application Layer
  43. Which internet layers do these protocols sit?

    TCP / UDP
    Transport Layer
  44. Which internet layer does this protocol sit?

    Internet Layer
  45. Which internet  layer does this protocol sit?

    Physical / Data Link Layer
  46. How many bits wide are internet addresses?
    32-bit wide
  47. How many classes of internet address are there? (And describe their structure)

    • Class A:  0 | net id | host id
    • Class B:  1 | 0 | net id | host id
    • Class C:  1 | 1 | 0 | net id | host id
    • Class D:  1 | 1 | 1 | 0 | multicast address
    • Class E:  1 | 1 | 1 | 1 | 0 | reserved for future
  48. Explain what dotted decimal notation is and describe how it is used to represent internet addresses.
    • Take an internet address
    • Separate it into blocks of 8 bits
    • Convert each block to its decimal equivalent
    • Piece back together but separate each block with a dot.
  49. Name the three special reserved addresses for internet addresses
    • All 1's - Local network broadcast
    • Net ID + all 1's - Directed broadcast for net
    • 127 + anything - Loopback (Which should never appear on the network)
  50. What does ARP stand for?
    Address Resolution Protocol
  51. What does ARP do?
    Maps an IP address onto Physical address

    • Takes an ARP request from the host which contains the target's IP address.
    • Target responds with and ARP reply which contains the physical address.
  52. What does the Internet Protocol do?
    • Defines the Internet Datagram
    • Defines the routing function
    • Specifies rules for datagram processing by hosts and routers
  53. When does Datagram Fragmentation occur?
    When the maximum network frame size is smaller than the maximum datagram size.
  54. Explain datagram fragmentation
    • When a datagram is too large for the network medium, it must be fragmented and sent in fragments. 
    • Each fragment has a header which denoted the order it belongs in the complete set
    • The fragments are reassembled at the destination and not at intermediary routers!
  55. What is the User Datagram Protocol (UDP)?
    • Enables datagrams to be sent and received from processes within a machine (using port numbers to define process)
    • Provides unreliable datagram transport
    • Is a thin transport protocol
  56. What does the Transmission Control Protocol provide?
    • Reliable stream delivery of ordered data
    • Virtual circuit connections
    • Full duplex connections
  57. What does ICANN stand for and do?
    • Internet Corporation for Assigned Names and Numbers
    • Allocates IP address space
    • Protocol identifier assignment
    • Top-Level Domain name system management
    • Root server system management functions
  58. Which layer does the DNS operate over?
    Application Layer
  59. Roughly how many root servers are there in the world?
    Roughly a dozen
  60. Given the following stages, describe in which order a domain request travels in order to receive the IP address.

    Root name Server
    Requesting Host
    Local name Server
    Authoritative name server
    • The requesting host sends a request to the local name server
    • The local name server searches in its table for the IP. If found it returns, else it sends the request on to the root server.
    • The root server will retrieve the name from the authoritative name server and send it back down the chain to the requesting host.
  61. What does TELNET do?
    Provides service to remote machine which allows the user on the client side to interact with the operating system sitting on the server side.
  62. Which protocol does TELNET use to send messages between its client and server
  63. What is NVT?
    Network Virtual Terminal
  64. What is used to send commands between:

    NVT (Network Virtual Terminal) format
  65. What does FTP allow a user to do?
    Update / Delete / Add / Read / List remote files.
  66. What does the SNMP do?
    • Concerned with managing all communication protocols.
    •     - Fault management
    •     - Performance
    •     - Layer management
    •     - Name mangement
    •     - Security management
    •     - Accounting management
  67. On which layer does the HTTP operate?
    Application Layer
  68. Which protocol does HTTP use to send messages?
  69. How does Peer-to-peer networking work?
    • Each peer publishes a list of its shared files to the network
    • When another peer searches for a particular file, the network returns a list of peers that have it.
    • The peer chooses one or more peers from the list and connects to them
  70. Other than file sharing, what things can you do with P2P applications?
    • Communication
    • Play P2P games
    • Parallel Computation
  71. Describe the architecture of eDonkey
    • eDonkey is made up of servers and clients.
    • The servers contain a list of files that its connected clients have
    • The server is responsible for connecting clients to each other when a file is to be shared
  72. How does Gnutella work when it wants to find a file?
    • Peer P will send a QUERY message to all direct neighbours in the network
    • If a neighbour has the file, it responds with a QUERY RESPONSE message
    • Otherwise, the neighbour forwards the QUERY to its neighbours, and so on
  73. How does BitTorrent work?
    • Files are divided into pieces
    • A peer can download different pieces of the same file
    • Once the peer downloads a piece, that peer then re-uploads it themselves.
  74. Explain how a distributed hash table can work for storing files and peers in a P2P network
    • Distributed hash tables should make a hash value of the PeerID and the FileID. 
    • A file should be stored with a peer who has the closest PeerID to the FileID.
    • When a user wishes to search for a file, they start at a relatively close PeerID (to the FileID) and iterately move closer toward the Peer with the closest PeerID
  75. P2P systems classifications:
    Define the four Centralisation degrees.
    • Class1: Purely decentralised: All nodes execute exactly the same tasks
    • Class2: Partially decentralised: Some peers have more important roles
    • Class3: Distributed servers: The system contains distinguished clients and servers
    • Class4: Centralised: One central server is designated for all or each of the crucial tasks
  76. P2P systems classifications:
    Define the two Structure degrees.
    • Unstructured: Nodes are strictly autonomous and store only information concerning their own files.
    • Structured: The system imposes storing a distributed information about the files locations.
  77. What is a web service?
    A software system designed to support interoperable machine-to-machine interaction over a network