Outline

• TCP Congestion-Avoidance
• The Application Layer
• DNS
• E-mail
• More on HTTP

Congestion-Avoidance Mechanisms

• As described so far, TCP needs to cause congestion in the network to best find the way to consume network resources.
• Alternatively, we might want to achieve the effect of using our network resources efficiently, but avoid congestion altogether.
• Such a strategy is called a congestion-avoidance strategy.

DECBit

• This was an early mechanism to do collision avoidance, developed by DEC.
• In this scheme each router monitors the load it is experiencing and explicitly notifies the end nodes when congestion is about to occur.
• This is done by setting a binary congestion bit in the packets that flow through the router. Hence the name DECBit.
• The destination host copies this bit into the ACK and when this ACK is seen by the sender it adjusts its window accordingly.
• More specifically, a router sets its DECbit if its average queue length is greater than or equal to 1.
• A sending increases its queue size by 1 if few than 50% of the acks it received for the last windows worth of data had the DECbit on.
• Otherwise, it decreases its window size by 0.875.

Random Early Detection (RED)

• In this scheme a router maintain a running average of its queue lengths.
• It also maintains two variables MinThreshold and MaxThreshold.
• If a packet arrives and the average queue length is less than MinThreshold it just queues the packet.
• If its between MinThreshold and MaxThreshold it drops the packet with probability p
• Otherwise, it drops the packet.
• The second case, would cause the sender to timeout earlier than if RED was not used, and so if TCP was being employed the CongestionWindow would stabilize to a workable value without putting as much load on the network.
• When RED is used, the probability p is typically a linear function of where between MinThreshold and MaxThreshold the average length currently is.

The Application Layer

• For the rest of today, we will talk about the top layer in the OSI model: the Application Layer.

The Domain Name System (DNS)

• To refer to a process on the internet we need to give an IP address and a port.
• These numbers on not very convenient for people to use, ASCII names would be better.
• An Arpanet scheme to solve this problem was to have a file hosts.txt in which host names in ASCII together with IP addresses were listed. Then if a connection was attempted to the ASCII name, a look up of the IP address was done.
• As the internet grew it became impractical for each machine to maintain a complete list of all hosts.
• Instead a hierarchichal system called the Domain Name System (DNS) was developed to solve this problem.
• To map a host name onto an IP address an application layer program calls a library procedure known as a resolver (for example, InetAddress method getHostAddress() in Java).
• The resolver sends UDP packets to a local DNS server (Port 53), which then looks up the name and address and returns the IP address to the resolver.

More on DNS

• Domain Names are organized into a hierarchy.
• One has top level domains which include com, edu, gov, int, mil, net, org, biz, info, name, and pro and one has country names such as ca.
• Second level domain names have the form: some string, period, followed by a top level name. For example, my-company.com
• Then one has third level domains such as my-dept.my- company.com
• In principle, you can keep going but for most US companies we only consider three levels. For foreign ones you might have four:
  www.amazon.co.uk
• Names are case insensitive
• Each domain controls how it allocates names under it. For example, .co.uk is a .com in the UK.

Still More on DNS

• Every domain can have a set of resource records associated with it.
• The most common resource record is just its IP address, but many other resource records also exist.
• When a resolver gives a domain name to a DNS server what it gets back are the resource records associated with the name.
• A resource records has the format:
  Domain_name Time_to_live Class Type Value
• This is the domain name to which the record applies, how many seconds the record is good for, for internet information the class will always be IN, and the type says what kind of records it is: SOA -(start of authority) parameters for this zone, A - IP address, MX - mail exchange, NS- name server of the domain, CNAME -- domain name (used for aliases), PTR -- alias for an IP address, HINFO -- host CPU and OS, TXT -- anything the register wanted to add
• It is completely possible for the domain name to be the same for an A record and an MX records, but to have different IP addresses. So the mail server for that entry would be different than where the rest of the traffic would go to.

Name Servers

• The DNS name space is divided into non-overlapping zones.
• Each zone contains part of the name tree and has name servers holding information about that zone.
• A zone will typically have a primary name server which gets information about that zone on disk, as well as one or more secondary servers.
• Zone boundaries are set-up by the given zone's administrators.
• When a host does a DNS lookup, it first checks its cache, if it misses it contacts the local DNS server, which in turn acts as a resolver as follows.
• When the local DNS server/resolver has a query it looks in its local cache, if the cache misses, it asks a root name server.
• The resolver then gets the name server for the domain requested (the NS record). It also gets the A record corresponding to this NS record. It then queries that to find the name server for the subdomain, and so on, until the name is finally resolved.

E-mail

• E-mail is distributed using a collection of application layer protocols:
  • SMTP (for sending e-mail between servers). The typical thing to imagine is that Server A has a set of users and Server B has a set of users. Then a mail message from a user on A can be sent to one on B, and put into a "drop-box" there. SMTP can also be used to relay messages. When Server A and B cannot directly communicate, but Server A knows Server C is allowed to talk to B and is willing to relay, then you could go from A to C to B.
  • POP3, IMAP for reading e-mail. These are used from a client on a machine different than the mail server and allow a user of a Server to read and manipulate the contents of their drop-boxes.
• sendmail and postfix are some examples of open-source SMTP servers.
• courier and dovecot is an example of mail transfer agent which supports POP3 and IMAP as well as several other protocols.

SMTP

• Stands for Simple Mail Transport Protocol.
• It uses port 25.
• A sample session with an SMTP server might look like:

S: 220 myname.edu
C: HELO client.domain 
S: 250 Hello client.domain, pleased to meet you 
C: MAIL FROM: bob@client.domain 
S: 250 bob@client.domain ... Sender ok 
C: RCPT TO: sally@myname.edu 
S: 250 sally@client.domain ... Receiver ok 
C: DATA 
S:354 Enter mail, end with "." on a line by itself 
C: How about lunch? 
C: . 
S: 250 Message accepted for delivery 
C: QUIT
S: 221 myname.edu closing connection

Format of Mail Messages

• E-mail messages consist of a header followed by a blank line followed by the contents of the message.
• A E-mail header consists of a sequence of lines the form something: value .
• A header must have a From: and To: line.
• An example might be

  From: bob@client.domain 
  To: sally@client.domain 
  Subject: lunch?
  How about lunch?
  

• For non-ASCII data one can specify MIME header -- Multipurpose Internet Mail Extension header, such as Content-Type: and Content-Transfer- Encoding:
• The receiver server will often append to the header Received: lines saying when the letter got to the server and whether it was forwarded through any other servers.

Additional SMTP Command

• Although all server support HELO, if you want to signal to the server that you are a more modern sender of mail, you use:

  ELHO client.domain
  

  instead.
• To use encryption, before issuing the HELO/ELHO command, a client can issue

  STARTTLS
  

  command.
• Many email servers only accept mail when encryption is used. If, as a human, you want to try to send mail to such a machine, you can use:

  openssl s_client -starttls smtp -crlf -quiet -connect pollett.org:25
  
  or
  
  openssl s_client -starttls smtp -crlf -quiet -connect pollett.org:587
  

  Here 25 and 587 are port numbers. 587 usually defaults to secure mail. Once connected you can issue usual SMTP commands.
• The flag -debug can be added to the above if you want to see a complete hex dump of the transmissions that are sent back and forth in addition to the text.
• If you need to authenticate before sending mail for some server, you can try to use plain authentication or login authentication:

  C: AUTH LOGIN
  S: 334 some_base64_string
  C: my_base_64_login_for_machine
  S: 334 some_other_base64_string
  C: my_base_64_password_for_machine
  S: 235 Authentication successful
  

POP3

• POP (Post Office Protocol) is a very simple protocol for a mail client to communicate with a mail server to find retrieve the mail for a particular user.
• A POP server listens to port 110.
• A client connecting to such a server might do:

  telnet mailserver 110 
  +OK POP3 server ready 
  user bob
  +OK 
  pass secret 
  +OK user successfully logged on
  

• The client can then issue commands such as:
  • list - which returns what messages are available and how big they are;
  • retr n -- retrieves message n,
  • dele n -- deletes message n, and
  • quit -- ends the session.

IMAP

• IMAP (Internet Mail Access Protocol) is a more sophisticated protocol which has the same role as POP3.
• Beyond the functionality of POP, it supports marking messages as read, multiple folders, moving messages between folders.

HTTP

• HTTP stand for hypertext transfer protocol
• One can connect to a web server using HTTP as follows:

  C: telnet www.cs.sjsu.edu 80
  S: Trying 130.65.86.46... 
     Connected to www.cs.sjsu.edu. Escape character is '^]'.
  C: GET /index.html HTTP/1.1 
     Host: www.cs.sjsu.edu
     [blank line]
  S: HTTP/1.1 200 OK 
     Date: Thu, 10 May 2007 18:37:34 GMT 
     Server: Apache/2.2.2 (Fedora)
     Last-Modified: Mon, 07 May 2007 22:38:22 GMT 
     Accept-Ranges: bytes
     Content-Length: 9700
     Connection: close
     Content-Type: text/html 
     [blank line]
     document
  

• The client always sends the first message.
• Besides the command GET (which requests a web document), a client can also issue commands such as POST, HEAD, etc.
• After the command line, the client can provides additonal lines of the form "field:value". If HTTP/1.1 is being used at least one of these fields must be the Host field.
• Then the client has a blank line and the server can start its response.
• The first line back says a status code for the response. In the above HTTP/1.1 200 OK means the request was successful
• Then, as with client, the server can have one or "field:value", followed by a blank line followed by any document data that will be sent back.