EEE3080F Communication Network and System Fundamentals

Transcription

1 EEE3080F Communication Network and System Fundamentals H Anthony Chan; Yang Li h.a.chan@ieee.org; yli@crg.ee.uct.ac.za Department of Electrical Engineering University of Cape Town Application layer Principles of network applications Web and HTTP FTP Electronic Mail: SMTP, POP3, IMAP Ssh DNS Port numbers P2P file sharing Socket programming with TCP Socket programming with UDP 204 Page 1 February 23, Page 2 February 23, 2007 The Web: the http protocol http: hypertext transfer protocol Web s application layer protocol / model : browser that requests, receives, displays Web s : Web sends s in response to requests http1.0: RFC 1945 http1.1: RFC 2068 PC running Explorer Mac running Navigator Server running NCSA Web The http protocol http: TCP transport service: initiates TCP connection (creates socket) to, port 80 accepts TCP connection from http messages (application-layer protocol messages) exchanged between browser (http ) and Web (http ) TCP connection closed http is stateless maintains no information about past requests Aside: Protocols that maintain state are complex! past history (state) must be maintained if / crashes, their views of state may be inconsistent, must be reconciled 204 Page 3 February 23, Page 4 February 23, 2007

2 http 1.0 example user enters URL, containing text, references to 10 jpeg images http example (cont.) 1a. http initiates TCP connection to http (process) at Port 80 is default for http. 2. http sends message (containing URL) into TCP connection socket 1b. http at host waiting for TCP connection at port 80. accepts connection, notifying 3. http receives request message, forms response message containing requested (somedepartment/home.index), sends message into socket 5. http receives response message containing html file, displays html. Parsing html file, finds 10 referenced jpeg s 6. Steps 1-5 repeated for each of 10 jpeg s 4. http closes TCP connection. 204 Page 5 February 23, Page 6 February 23, 2007 Non-persistent, persistent connections http message format: request Non-persistent http/1.0: parses request, responds, closes TCP connection 2 RTTs to fetch TCP connection request/transfer each transfer suffers from TCP s initially slow sending rate many browsers open multiple parallel connections Persistent default for htp/1.1 on same TCP connection:, parses request, responds, parses new request,.. sends requests for all referenced s as soon as it receives base HTML. fewer RTTs, less slow start. two types of http messages: request, response message: ASCII (human-readable format) request line (GET, POST, HEAD commands) header lines GET /somedir/page.html HTTP/1.0 User-agent: Mozilla/4.0 Accept: text/html, image/gif,image/jpeg Accept-language:fr 204 Page 7 February 23, 2007 Carriage return, line feed indicates end of message (extra carriage return, line feed) 204 Page 8 February 23, 2007

3 message: general format http message format: response method sp URL header field name : header field name : cr lf sp version value cr lf value cr lf Entity Body cr lf request line header lines status line (protocol status code status phrase) data, e.g., requested html file header lines HTTP/ OK Date: Thu, 06 Aug :00:15 GMT Server: Apache/1.3.0 (Unix) Last-Modified: Mon, 22 Jun Content-Length: 6821 Content-Type: text/html data data data data data Page 9 February 23, Page 10 February 23, 2007 status codes in first line in response message A few example codes: 200 OK request succeeded, requested later in this message 301 Moved Permanently requested moved, new location specified later in this message (Location:) 400 Bad Request request message not understood by 404 Not Found requested document not found on this 505 HTTP Version Not Supported Trying out http ( side) for yourself 1. Telnet to your favorite Web : telnet Type in a GET : GET /~ross/index.html HTTP/1.0 Opens TCP connection to port 80 (default http port) at Anything typed is sent to port 80 at By typing this in (hit carriage return twice), you send this minimal (but complete) GET request to http 3. Look at response message sent by http! 204 Page 11 February 23, Page 12 February 23, 2007

4 User- interaction: authentication Cookies: keeping state Authentication: control access to content authorization credentials: typically name, password stateless: must present authorization in each request authorization: header line in each request if no authorization: header, refuses access, sends WWW authenticate: header line in response 401: authorization req. WWW authenticate: + Authorization: <cred> usual msg + Authorization: <cred> usual msg -generated #, -remembered #, later used for: authentication remembering user preferences, previous choices sends cookie to in response msg Set-cookie: presents cookie in later requests cookie: usual + Set-cookie: # cookie: # usual msg cookie: # usual msg cookiespectific action cookiespectific action 204 Page 13 February 23, Page 14 February 23, 2007 Conditional GET: -side caching Goal: don t send if has up-to-date cached version : specify date of cached copy in If-modified-since: : response contains no if cached copy is up-to-date: HTTP/ Not Modified msg If-modified-since: HTTP/ Not Modified msg If-modified-since: HTTP/ OK <data> not modified modified Web Caches (proxy ) Goal: satisfy request without involving user sets browser: Web accesses via web cache sends all http requests to web cache in web cache: web cache returns else web cache requests from, then returns to Proxy 204 Page 15 February 23, Page 16 February 23, 2007

5 Assume: cache is close to (e.g., in same network) smaller response : cache closer to decrease traffic to distant s link out of institutional/local ISP network often bottleneck Why Web Caching? institutional network public Internet 1.5 Mbps access link 10 Mbps LAN s HTTPS HTTPS also transfers data using HTTP. However, instead of using plain text socket communication, the session data is encrypted using a version of the Secure Socket Layer (SSL) or Transport Layer Security (TLS) protocols, thus ensuring reasonable protection from eavesdroppers and man in the middle attacks. The default TCP port of an https: URI is 443 (for unsecured HTTP, the default is 80). institutional cache 204 Page 17 February 23, Page 18 February 23, 2007 WebDAV (Web Distributed Authoring and Versioning): WebDAV or DAV is a set of extensions to the HTTP/1.1 protocol (adding new HTTP methods and headers) which allows users to collaboratively edit and manage files on remote web s provides authoring support for Web resources of any media type (HTML, GIF, JPEG, etc.) enjoys the HTTP benefits of strong authentication, encryption, proxy support, and caching. can pipeline multiple file transfers through a single TCP connection (RFC 2518) EEE3080F Communication Network and System Fundamentals God made the world rich enough to feed and clothe all Page 19 February 23, Page 20 February 23, 2007