Project 4: Internet Router

Due 04/17/05

Educational Objectives: Experience with hash tables, IP addresses, and Internet routing. Experience creating both servers and clients for hash tables.

Operational Objectives: Implement the template adaptor classes THashTable<K,T,H> and THashTableIterator<K,T,H>. Design and implement a class RouteTable to be used with the distributed client program iprouter.cpp to simulate the operation of an Internet router.

Deliverables: Three files:

thashtbl.h   # contains THashTable<> and THashTableIterator<> template classes (including implementations)
iptbl.h      # contains RouteTable class definition and supporting function prototypes
iptbl.cpp    # contains RouteTable implementations and supporting function implementations

IP Addressing

There are two ways to represent ip addresses: the 4-number "dot" notation and the 32-bit (4-byte) "number" notation. The dot notation we store as a String object (typedef ipString) and the number notation we store as an unsigned long object (typedef ipNumber). Generally, a router uses ipNumber as its internal representation, while externally across the Internet the ipString representation is used.

The ipString representation consists of four numerical fields separated by (three) dots. For example, 128.186.121.211 is the ip address of a machine in the computer science department. Each numerical field in this address represents a number in the range [0, 255]. (Numbers 256 or greater make an invalid ipString. We also define the zero address 0.0.0.0 as invalid.) This number in turn denotes an 8-bit (one byte) quantity. The four numerical fields, concatenated, represent 32 bits or 4 bytes. This 4-byte number is the internal ipNumber representation. In general, we use hexadecimal (base 16) representation to denote ipNumber objects. The ipNumber representation of the address above is 10000000101110100111100111010011 (bin) = 80BA79D3 (hex).

The ipClass of an ip address is defined in terms of the bits in its ipNumber representation. There are three recognized classes of ipNumbers: A, B, and C. An ipNumber that is not one of these classes is called bad and cannot be used. Class A consists of ipNumbers beginning (on the left) with bit '0'. Class B begins with bits '10'. Class C begins with bits '110'. All other ipNumbers are bad. (Note that bad ipNumbers are not the same as invalid ipStrings.) A good (i.e., not bad) ipNumber contains information in fields called netID and hostID. The ranges of these fields depend on the class, as shown in the following table (numbering the bits from the left, starting with 1):

Class A:            netID = bits 2..8;            hostID = bits   9..32 
Class B:            netID = bits 3..16;           hostID = bits  17..32
Class C:            netID = bits 4..24;           hostID = bits	 25..32 

The netID and hostID are full 32-bit words with the irrelevant bits masked to zero. For example, the address 80BA79D3 is class B with netID = 00BA0000 and hostID = 000079D3. netID and hostID are used by a router to forward messages. We do not express netID and hostID in "dot" notation. The following summarizes the calculations:

ipString            128.186.121.211
ipNumber  (bin)     1000 0000 1011 1010 0111 1001 1101 0011
ipNumber  (hex)     8    0    B    A    7    9    D    3    = 80BA79D3
ipClass             class B
netMaskB  (bin)     0011 1111 1111 1111 0000 0000 0000 0000
netMaskB  (hex)     3    F    F    F    0    0    0    0    = 3FFF0000
netID     (bin)     0000 0000 1011 1010 0000 0000 0000 0000
netID     (hex)     0    0    B    A    0    0    0    0    = 00BA0000
hostMaskB (bin)     0000 0000 0000 0000 1111 1111 1111 1111
hostMaskB (hex)     0    0    0    0    F    F    F    F    = 0000FFFF
hostID    (bin)     0000 0000 0000 0000 0111 1001 1101 0011
hostID    (hex)     0    0    0    0    7    9    D    3    = 000079D3

Real Routers

Routers are special-purpose computers that monitor Internet traffic, either rejecting or accepting messages. Accepted messages are forwarded and rejected messages are ignored. For example, the Love Building router run by the Computer Science department rejects incoming messages that are not addressed to a machine in the building, and accepts and forwards messages addressed to one of the machines in the building. It also forwards outgoing messages to another router for further routing. Much of the computation has direct hardware support, enabling speed (bandwidth) to be 1 gigabit. The route tables, in contrast, need to be software tables so that the router can be programmed as both internal and external machine configurations and addresses evolve. Hash tables/maps are the only data structure that guarantees fast enough lookup for today's high-bandwidth routers.

Procedural Requirements

  1. Create and work within a separate subdirectory cop4530/proj4.

  2. Do your own work. Variations of this project have been used in previous courses. You are not permitted to seek help from former students or their work products. For this and all other projects, it is a violation of course ethics and the student honor code to use, or attempt to use, code from any source other than that explicitly distributed in the course code library, or to give or receive help on this project from anyone other than the course instruction staff. See Introduction/Work Rules.

  3. Begin by copying the following files from the course directory /home/courses/cop4530p/spring05 into your proj4 directory:

    LIB/proj4/iprouter.cpp
    LIB/proj4/thashtbl.partial
    LIB/proj4/makefile
    LIB/proj4/iptbl.?
    LIB/area54/iprouter_s.x
    LIB/area54/iprouter_i.x
    

    The naming of these files uses the convention that _s and _i are compiled from the same cource code on program (Sun/Solaris) and linprog (Intel/Linux), respectively. The executables are distributed only for your information and experimentation. You will not use these files in your own project.

  4. You are to define and implement the template classes THashTable<K,T,H> and its associated iterator class THashTableIterator<K,T,H>. In addition you are to design and implement the class RouteTable to be used with the distributed client program iprouter.cpp.

  5. File thashtbl.h should contain the definitions and implementations of the template classes THashTable<K,T,H> and THashTableIterator<K,T,H>.

  6. File iptbl.h should contain the definition of class RouteTable and prototyes of supporting functions.

  7. File iptbl.cpp should contain implementations of RouteTable methods and supporting functions.

  8. A makefile is supplied that (1) compiles your project components into separate object code files, (2) builds an executable iprouter.x from these object code files, and (3) builds an executable fthtbl.x directly from the library file tests/fthtbl.cpp. You can test different targets individually by nameing them as a parameter to the make command.

  9. Copy the file LIB/submitscripts/proj4submit.sh into your project directory, change its permissions to executable, and submit the project by executing the script.

    Warning: Submit scripts do not work on the program and linprog servers. Use shell.cs.fsu.edu to submit projects. If you do not receive the second confirmation with the contents of your project, there has been a malfunction.

Code Requirements and Specifications - THashTable and THashTableIterator

  1. Implement the THashTable<K,T,H> and THashTableIterator<K,T,H> classes as defined in the file LIB/proj4/thashtbl.partial using the implementation plan discussed in the lecture notes.

  2. Place THashTable<K,T,H> and THashTableIterator<K,T,H> in the fsu namespace.

  3. For the private data storage use a fsu::TVector < fsu::TList < fsu::TEntry < key_type, data_type > > > object. (This is implied by the definition in thashtbl.partial.)

  4. Use the following C = fsu::TList operations, where T is C::value_type:

    C::Iterator Includes (const T& t);                 // returns location of t if found
    int         Insert   (C::Iterator& I, const T& t); // places copy of t at I
    int         Remove   (C::Iterator I);              // removes item at I
    int         Empty    ();                           // true iff Size() returns zero
    size_t      Size     ();                           // returns the number of elements in C
    C::Iterator Begin    ();                           // returns iterator to first element of C
    C::Iterator End      ();                           // returns iterator past the last element of C
    

    Note this means that THashTable:: search, insert, and remove operations should all use C::Includes(t) operation followed by a constant time insert or remove operation as appropriate. (TList::Includes is an implementation of sequential search.)

  5. Be sure not to change the definition of THashTable<> from that distributed in LIB/proj4/thashtbl.partial.

  6. Place all hash table code, including definitions and implementations, in the file thashtbl.h.

  7. Thoroughly test your implementation for correct functionality using the provided test client tests/fthtbl.cpp and tables from tests/tables/ and/or tables generated using tests/rantable.cpp

Code Requirements and Specifications - RouteTable

  1. RouteTable Types. You will need the following type definitions:

    typedef unsigned long ipNumber;
    typedef String        ipString;
            
    enum ipClass 
    {
      classA, classB, classC, badClass
    } ; 
    
  2. Output operator.Overload operator <<() for type ipClass using the following prototype:

    std::ostream& operator << (std::ostream& os, ipClass ipc);
    // sends 'A', 'B', 'C', or 'D' to os depending on ipClass value
    
  3. RouteTable Public Interface. You will need the following public methods in RouteTable:

    void Load        (const char* tblfile);
    void Save        (const char* tblfile);
    void Insert      (const ipString& destination, const ipString& route);
    void Remove      (const ipString& destination);
    void Go          (const char* msgfile, const char* logfile);
    void Clear       ();
    void Dump        (const char* dumpfile);    
         RouteTable  (unsigned int approx_size);
         ~RouteTable ();
    
    static ipClass ipInterpret(const ipNumber& address, ipNumber& netID, ipNumber& hostID); 
    // returns ipClass and sets netID and hostID of address; 
    // if address is badClass, netID and hostID are set to 0. 
    
    static ipNumber ipS2ipN (const ipString& S); 
    // converts ipString to ipNumber, checking for syntax and oversize errors 
    

    Go() performs a simulation as described.

    Load() and Save() build/save the route table from/to an external file.

    Insert() and Remove() are similar to the standard table operations, except that they must translate input from ipString to ipNumber prior to accessing the underlying table.

    Go() and Dump() send output to screen when passed a 0 pointer as output file parameter; otherwise all char* parameters are treated as external file names.

    Static member functions behave like stand-alone functions with scope limited to the class. These functions should be included in the RouteTable class as static methods.

  4. Data Structures. Use a THashTable<ipNumber,ipNumber,ipHash> object as the primary data structure supporting the route table.
    Note : This object will need to be created dynamically in order to set the number of buckets.

  5. Data, Table, and File Format. RouteTable objects use three distinct kinds of files: table files, message files, and log files.

    A table file (extension .tbl) consists of pairs of ipNumber (one pair per line) written in hex notation with '0' fill, so that all entries have 8 characters. The Load(filename) method of a RouteTable object reads data from a table file and inserts the data into its internal table. The Save(filename) method writes all data in the internal table to a table file.

    The internal table also uses ipNumber representation of addresses.

    The ipNumber pairs in a route table are destination and route pairs. The destination is the key, and the route is the data in the table.

    When individual entries are inserted, removed, or looked up (via the RouteTable public interface) in the internal table, the external ipString representation is used for input.

    The Go() method is the method that simulates a router in operational mode. The incoming internet traffic is simulated by a message file. Each line of the message file can be thought of as a "packet" with a destination and a message body. The result of routing this traffic is recorded in a log file.

    Go() should read message packets (lines) one at a time from a message file and write the disposition of these message packets to a log file. For each message packet (line of the message file), the router (1) reads the destination address and converts it to ipNumber form; (2) rejects the message packet if the ipClass is bad; (3) looks up the packet destination in the internal table; (4) if not found, rejects the packet; (5) routes the packet using the netID and hostID of the route retrieved from the table; (6) writes the disposition of the message packet to the log file.

    The message file (extension .msg) consists of two strings per line. The first string of the packet (line) is an ip address in dot notation (intended to be the destination of the message) and the second is a message ID (simulating the body or content of a real message). Each line of this file represents one "message packet" coming in to the router.

    The log file (extension .log) should contain one line for each message packet in the message file. The log entry for a message should begin with the message ID, followed by the ipNumber of the destination (8-digit hex), the ipClass, and the routing information for the message.

  6. Running a Simulation. Your RouteTable should work with the supplied client program iprouter.cpp, which simulates the operation of a router.

Hints