Transcript 08-Application File

Application Layer Protocols
Application Layer Protocol
Transport Service
Mail Protocols
TCP
Hyper-Text Transfer Protocol
TCP
File Transfer Protocol
TCP
Telnet
TCP
Trivial File Transfer Protocol
UDP
Network File System
UDP
Domain Name System
UDP
Simple Network Management Protocol
UDP
2/11/10
08-Application
1
Mail Protocols
• Simple Mail Transfer Protocol (SMTP) - for transmitting
e-mails to a server and between servers
• Post Office Protocol - Version 3 (POP3) – for retrieving emails from a server
• Internet Mail Access Protocol (IMAP) – alternative
improved protocol for retrieving e-mails
• Multipurpose Internet Mail Extensions (MIME) –
supports different formats for data attached to e-mails
• ISO equivalent protocol is Message-Oriented Text
Interchange Standard (MOTIS), also known as X.400 or
Message Handling System (MHS)
2/11/10
08-Application
2
E-Mail Systems
• E-mail systems involve user agents and
message transfer agents (mail servers)
• E-mail systems support 5 basic functions:
–
–
–
–
–
2/11/10
Composition
Transfer
Reporting
Displaying
Disposition
08-Application
3
SMTP
• Originally defined in RFC 822 in 1982
• Simple ASCII text based messaging protocol
• Restricted to 7 bit ASCII code – binary data needs to
be encoded in ASCII
• Non-real time store and forward protocol
• Messages comprise a header and a body
• SMTP along with all other e-mail protocols requires
a reliable transport service and runs over TCP
2/11/10
08-Application
4
RFC 822 Headers Fields
related to Message Transfer
© Tanenbaum, Prentice Hall International
2/11/10
08-Application
5
Other RFC 822 Headers Fields
not related to Message Transfer
© Tanenbaum, Prentice Hall International
2/11/10
08-Application
6
MIME
• Multipurpose Internet Mail Extensions (MIME)
– MIME is defined in RFC 2045
– It allows non-ASCII data, such as images, audio, video and
8 bit application data to be sent via e-mail
– Data has to be encoded into ASCII and decoded at the
receiving end
– The type of data is indicated in a MIME header so that the
user agent knows which application to use when asked to
display/process the data
– Mime data types are used by most other Internet
applications
2/11/10
08-Application
7
RFC 822 Header Fields
added by MIME
© Tanenbaum, Prentice Hall International
2/11/10
08-Application
8
Some MIME Types and Sub-types
type/sub-type
Description
text/plain
text/html
image/gif
image/jpeg
application/msword
application/pdf
audio/basic
video/mpeg
Unformatted text
HTML format
Still picture in GIF format
Still picture in JPEG format
Microsoft Word format
Acrobat Portable Document Format
Audible sound
Video in MPEG format
2/11/10
08-Application
9
POP3
• Simple but limited mail access protocol
• 3 phases
– Authentication (username, password)
– Transaction (user agent retrieves messages)
– Update (server deletes messages, if required, after
retrieval completed)
• User Agent acts as a client to the Mail Server
– 4 requests
list, retr, dele, quit
– 2 responses
+OK, -ERR
2/11/10
08-Application
10
IMAP
• Supports same functions as POP3, but has more
features and is more complex
• It allows mail folders to be stored on the server
• IMAP is designed to support nomadic workers
who want to access their e-mail from different
machines
• It allows users to retrieve headers, message bodies
and individual parts of multipart MIME messages,
thus conserving bandwidth on a low bandwidth
connection
2/11/10
08-Application
11
Web-based E-mail
• Introduced by Hotmail in the mid 1990s
• User agent is a browser which retrieves
messages from and sends messages to a
Web Server using HTTP
• Provides enormous convenience as all the
user needs to access mail is a browser and
an internet connection
• Like IMAP messages can be organised into
folders on the Server
2/11/10
08-Application
12
HTTP (Hyper-Text Transfer Protocol)
• Client server protocol defined in RFC 1495 and RFC
2616 that allows a user agent (browser) to view web
pages stored on Servers
• A web page consists of a collection of objects stored
in files on the server, such as HTML files, JPEG
files, GIF files and Java applets. The main HTML
file references the other objects to be
displayed/processed using URLs
• HTTP defines how the browser requests web pages
from the server and how the server transfers web
pages to the browser
2/11/10
08-Application
13
Uniform Resource Locators
• URLs are used by browsers to reference web
pages and other resources. They consist of three
parts, usually of the form:
protocol://hostname/resourcename:port number
• In terms of web pages, the URL determines how
the page can be accessed, where the page is
located and what the page is called. Valid
protocols are http, ftp, file, news, gopher, mailto or
telnet. These are all supported in the address
window of browsers
2/11/10
08-Application
14
HTTP Connections
• HTTP requires a reliable transport service and runs
over TCP connections.
• Connections can be
– Non-persistent (HTTP/1.0)
• Only one object is transferred and the connection is then closed and a
new transport connection opened to retrieve other referenced objects
– Persistent (default, but configurable in HTTP/1.1)
• Multiple objects can be transferred over the same transport
connection which is closed after a timeout. This improves response
times
2/11/10
08-Application
15
HTTP Pipelining
• Persistent Connections can also support
pipelining
– Pipelining further improves response times by
allowing the client to issue new requests before
responses are received to earlier requests
– Default mode of HTTP/1.1 is persistent with
pipelining
2/11/10
08-Application
16
HTTP Requests
• The HTTP Header has a Method field that
indicates the type of request.
• The most commonly used method is GET
which requests the header and the body of
the web page
• If only the header is required the HEAD
method is used
2/11/10
08-Application
17
HTTP Request Format
© Kurose & Ross, McGraw Hill
2/11/10
08-Application
18
HTTP Responses
• HTTP Responses include a status code and phrase
–
–
–
–
–
–
200 OK
301 Moved Permanently
400 Bad Request
401 Authorisation Required
404 Not Found
505 HTTP Version Not Supported
• The server will generate header lines in the
response appropriate to the version of HTTP
indicated in the request
2/11/10
08-Application
19
HTTP Response Format
© Kurose & Ross, McGraw Hill
2/11/10
08-Application
20
HTTP Authorisation
• Web Servers do not normally store any state
information about users (HTTP is said to be a
stateless protocol)
• Server responds with 401 Authorisation Required
and specifies how to authenticate in a header field
(E.g. username, password)
• Browser prompts user for username and password
• Browser resends the request and all subsequent
requests to this server in an Authorisation header
line
2/11/10
08-Application
21
HTTP Cookies
• Web Servers do not normally store any information
about their clients.
• Web Servers can keep track of users by storing
information, such as a user identifier on the client’s
hard drive. This information is called a cookie
• Web Servers store cookies with the Set-cookie
header line
• Clients transmit the cookie information whenever
they make a subsequent request to the server using
the Cookie header line
2/11/10
08-Application
22
HTTP Web Caching
• Web pages can be cached locally by the
browser and in proxy servers, to reduce
response times and inefficiency of
downloading recently accessed web pages
• Browser can make sure it has the latest
version of a page by using the If-modifiedsince header line with the time the page was
cached
2/11/10
08-Application
23
File Transfer Protocol (FTP)
• FTP is one of the original protocols designed for the
Internet. It requires a reliable transport service and
runs over TCP. It is used for uploading web pages
• It has a command line interface and supports many
complex functions
• It sets up two TCP connections, one for commands
and responses and the other for data transfer
• Anonymous FTP uses anonymous as a username and
by convention your e-mail address as the password
• ISO equivalent protocol is File Transfer Access and
Management (FTAM)
2/11/10
08-Application
24
Telnet
• Telnet is a virtual terminal protocol that provides
remote terminal access to a host. It runs over TCP
• It is again one of the original protocols designed for the
Internet
• Telnet can translate between different character codes
• It is extremely insecure. Usernames and passwords for
remote login are transmitted without encryption
• Firewall routers normally only forward telnet packets
from known IP addresses
• OSI equivalent protocol is Virtual Terminal (VT)
2/11/10
08-Application
25
Trivial File Transfer Protocol (TFTP)
• TFTP is a lightweight simple file transfer
protocol, designed for use over LANs with
low error rates
• It transmits one block of data and waits for
an acknowledgement before sending any
more data
• It does not require a reliable transport
protocol and runs over UDP
2/11/10
08-Application
26
Network File System
• NFS was developed by Sun Microsystems for Unix but
has been implemented for VMS, DOS, Windows and
MVS.
• It allows remote file systems (possibly using different
operating systems) to be accessed as if they were local.
• The physical location of directories and the remote
operating system is hidden from the users
• NFS makes use of Remote Procedure Calls (a simple
client server request/response) which is best suited to
running over UDP although it can be run over TCP in
some implementations
2/11/10
08-Application
27
Domain Name System (DNS)
• DNS is a Client Server Directory Service
for translating between host names and IP
addresses
• The protocol consists of short request and
response messages
• DNS does not require a reliable transport
service and runs over UDP
• ISO equivalent protocol is Directory Access
Protocol or X.500
2/11/10
08-Application
28
DNS Database
• The DNS is a hierarchical distributed database reflecting
the hierarchical structure of domains
• Allocation of names is delegated to the Zones who are
responsible for one or more domains who may delegate
it further. Each Zone runs at least two DNS Servers
• Top level domains are allocated by the Internet
Corporation for Assigned Names and Numbers
(ICANN)
• Top level domains also exist for each country
• ICANN coordinate 13 replicated root servers which
know the authoritative servers for each top level domain
2/11/10
08-Application
29
The Domain Hierarchy
© Tanenbaum, Prentice Hall International
2/11/10
08-Application
30
DNS Zones
© Tanenbaum, Prentice Hall International
2/11/10
08-Application
31
DNS Name Resolution
• The client (called the name resolver) will make a
request to its local DNS server. If this has the name
and address in cache it will return the IP address.
Note: addresses will time out in cache at an interval
determined by the administrator (typically 2 days)
• Otherwise a request will be made to a higher level
server or a root server which may have to refer the
request down to its sub-domain servers.
• The process will end either with an IP address being
returned or an error message to say that the address
could not be found
2/11/10
08-Application
32
Simple Network Management Protocol
• SNMP is a simple protocol for managing network
devices
• SNMP is often implemented in simple network
devices that cannot support a full and complex
protocol stack. It therefore uses UDP as a
transport service
• SNMP uses ASN.1 encoding
• The ISO equivalent protocol is the Common
Management Information Protocol
• SNMP will be studied in more detail in the second
half of the course
2/11/10
08-Application
33