CS 417 Exam 1
Spring 2008
- 4 points
Why do NUMA systems require operating system support whereas a bus or crossbar architecture does not? - 4 points
Your network requires the ability to address 33,000 hosts. A class B network, with a 16-bit host address, has the ability to address 65,534 hosts. 32,534 potential addresses are hence wasted. Will using CIDR instead of a class B address help? Explain how or, if not, explain why not. - 5 points
Why might it take a bit longer for a client to open a file under AFS than under NFS? - 6 points
(a) How does CIDR help us from running out of global IP addresses?
(b) How does NAT/PAT help us from running out of global IP addresses?
(c) How does IPv6 help us from running out of global IP addresses? - 5 points
How does AFS provide cache consistency across all clients with cached files? - 5 points
How do RPC systems make a remote function look and feel like a local function? Explain the steps undertaken in making a remote procedure call. - A snoopy cache eases congestion for which operations?
(a) Memory reads.
(b) Memory writes.
(c) Network receives.
(d) Network transmits. - Which statement best describes Java RMI's garbage collection?
(a) A client sends a dirty message to tell a server that a remote reference is used, periodically refreshes the dirty message, and finally sends a clean message when the object is no longer needed.
(b) The server periodically pings a client for a list of remote objects that it is referencing. The client responds with the list of objects. If it fails to respond or if an object is not in the returned list, the server then clears the references.
(c) Each time a client creates a reference to a remote object, it sends an increment message to the server. Each time the client removes a reference, it sends a decrement message. When the reference count goes to zero on the server, the object is cleared.
(d) An agent is sent from the server to each of the clients with a list of remote references on that server. If any of the references are no longer used by the client, the agent cleans them up. - TCP is an example of:
(a) a virtual circuit protocol.
(b) a connectionless protocol.
(c) a network layer protocol.
(d) all of the above. - Ethernet collision is a result of:
(a) two nodes on a LAN transmitting simultaneously.
(b) two nodes operating without synchronized clocks.
(c) two processes within a node attempting to send a packet at the same time.
(d) a node transmitting while a router is congested. - Which OSI layer is most likely buggy if you're having problems with recognizing integers and strings properly in a network message?
(a) transport (3)
(b) session (4)
(c) presentation (5)
(d) application (6) - Which system call associates a local IP address and port with a socket?
(a) socket
(b) bind
(c) listen
(d) connect - Compared to packet switching, circuit switching is likely to:
(a) provide higher bandwidth.
(b) result in a more efficient utilization of the network.
(c) provide a constant latency.
(d) incur more data errors . - DCE RPC improved on SUN (ONC) RPC by adding:
(a) an interface definition language.
(b) automatic generation of a unique identifier for the interface.
(c) an interface pointer to support objects.
(d) dynamic selection of TCP or UDP. - How is an IP address for a machine related to an ethernet address for that machine?
(a) They are the same if using the same network interface.
(b) The IP address is derived from the ethernet address.
(c) The ethernet address is derived from the IP address.
(d) They are not related. - How does UDP differ from TCP?
(a) UDP is a network layer protocol while TCP is a transport layer protocol.
(b) UDP is a connectionless protocol while TCP is connection-oriented.
(c) UDP can span multiple networks while TCP cannot.
(d) UDP doesn't require the use of port numbers while TCP does. - Traffic policing works by:
(a) discarding packets that exceed the allotted bandwidth.
(b) discarding packets that are not from a valid source or to a valid destination.
(c) queuing packets during peak loads for later retransmission.
(d) informing senders when they can or cannot transmit packets. - Hard QoS is most likely to employ:
(a) traffic prioritization.
(b) flow detection.
(c) the combination of multiple small packets into one big one.
(d) bandwidth reservation. - Before you can make a remote procedure call with Sun RPC, you need to contact the portmapper by calling the clnt_create function, This is an example of:
(a) static binding.
(b) early binding.
(c) late binding.
(d) delayed binding. - The Domain Name System is responsible for translating:
(a) human-readable computer names to IP addresses.
(b) IP addresses to ethernet MAC addresses.
(c) ethernet MAC addresses to their corresponding IP addresses.
(d) internal IP addresses to external IP addresses. - An interface definition language is a:
(a) way of specifying the names, parameters, and returns of remote procedures.
(b) language for coding remote procedures.
(c) language used for programming stub functions.
(d) definition of data encoding formats over the network. - DCE RPC enhanced Sun (ONC) RPC with:
(a) the ability to locate the server functions without knowing what server they're on.
(b) the ability to pass data structures in addition to basic data types.
(c) a pre-compiler to generate stubs.
(d) support for the instantiation of objects. - Microsoft created Microsoft RPC by enhancing DCE RPC by adding:
(a) an interface pointer identifier that identifies the interface for an object on a particular server.
(b) the ability to use HTTP GET, POST, PUT, and DELETE operations to perform requests.
(c) the ability to define the interface in a WSDL document.
(d) dynamic selection of TCP or UDP for communication. - SOAP does not help deal with:
(a) locating machines hosting remote services.
(b) creating requests that can cross firewalls more easily.
(c) different data encodings due to different machine architectures and programming languages.
(d) invoking methods on remote objects. - Which of the following is not a valid NFS function in the original version of NFS?
(a) lookup: look up a file name and return a handle to it.
(b) mkdir: create a directory.
(c) readbytes: read bytes from a file.
(d) close: close a file. - NFS does not improve performance via:
(a) read-ahead.
(b) larger-block reads.
(c) write-behind.
(d) long-term block caching.PART III – 10 points
For each statement, select whether it is true or false.
- A surrogate process in Microsoft DCOM is a server process that instantiates remote object requests.
[True] [False] - Because of its stateless architecture, NFS provides the user with session semantics for file access.
[True] [False] - A snoopy cache monitors memory activity on the system bus to keep its own cache coherent.
[True] [False] - Ethernet's CSMA/CD mechanism allows an ethernet controller to ensure that the packet went out without a collision.
[True] [False] - Nagle's algorithm reduces the number of TCP packets sent by adding new data to an unsent packet.
[True] [False] - ATM differs from IP in that it uses fixed-size packets while IP uses variable-size packets.
[True] [False] - To a client, a DCOM object just looks like a local COM object that happens to use RPC within it.
[True] [False] - CORBA's Internet Inter-ORB protocol (IIOP) defines a network protocol for communicating between clients and servers.
[True] [False] - A stateful file system is needed to support file locking.
[True] [False] - Ethernet is a broadband network since its speeds are generally greater than 10 Mbps.
[True] [False]
Part I – 30 Points
PART II – 60 points – 3 points each
For each statement, select the most appropriate answer.
