1

WEB ESSENTIALS: CLIENTS, SERVERS, AND COMMUNICATION THE INTERNET Technical origin: ARPANET (late 1960s) o One of earliest attempts to network heterogeneous, geographically dispersed computers Email first available on ARPANET in 1972 (and quickly very popular!) ARPANET access was limited to select DoD-funded organizations Open-access networks o Regional university networks (e.g., SURAnet) o CSNET for CS departments not on ARPANET NSFNET (1985-1995) o Primary purpose: connect supercomputer centers o Secondary purpose: provide backbone to connect regional networks Original NSFNET backbone speed: 56 kbit/s Upgraded to 1.5 Mbit/s (T1) in 1988 Upgraded to 45 Mbit/s (T3) in 1991 In 1988, networks in Canada and France connected to NSFNET In 1990, ARPANET is decommissioned, NSFNET the center of the internet Internet: the network of networks connected via the public backbone and communicating using TCP/IP communication protocol o Backbone initially supplied by NSFNET, privately funded (ISP fees) beginning in 1995 INTERNET PROTOCOLS Communication protocol: how computers talk Cf. telephone protocol: how you answer and end call, what language you speak, etc. Internet protocols developed as part of ARPANET research ARPANET began using TCP/IP in 1982 Designed for use both within local area networks (LANs) and between networks IP is the fundamental protocol defining the Internet (as the name implies!)

IP address: 32-bit number (in IPv4) Associated with at most one device at a time (although device may have more than one) Written as four dot-separated bytes, e.g. 192.0.34.166 IP function: transfer data from source device to destination device IP source software creates a packet representing the data Header: source and destination IP addresses, length of data, etc. Data itself If destination is on another LAN, packet is sent to a gateway that connects to more than one network

IP
Source Network 1

Gateway

Destination
Gateway

Network 2

Network 3

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

10

IP
Source LAN 1

Gateway

Destination
Gateway

Internet Backbone

LAN 2

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

11

TRANSMISSION CONTROL PROTOCOL (TCP) Limitations of IP: No guarantee of packet delivery (packets can be dropped) Communication is one-way (source to destination) TCP adds concept of a connection on top of IP Provides guarantee that packets delivered Provide two-way (full duplex) communication

TCP
Establish connection.

{
{ {
Source

Can I talk to you?

OK. Can I talk to you? OK.

Send packet with acknowledgment.

Heres a packet. Destination Got it. Heres a packet.

Resend packet if no (or delayed) acknowledgment.

Heres a resent packet.

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

TCP also adds concept of a port TCP header contains port number representing an application program on the destination computer Some port numbers have standard meanings Example: port 25 is normally used for email transmitted using the Simple Mail Transfer Protocol (SMTP) Other port numbers are available first-come-first served to any application

USER DATAGRAM PROTOCOL (UDP) Like TCP in that: Builds on IP Provides port concept Unlike TCP in that: No connection concept No transmission guarantee Advantage of UDP vs. TCP: Lightweight, so faster for one-time messages DOMAIN NAME SERVICE (DNS) DNS is the phone book for the Internet Map between host names and IP addresses DNS often uses UDP for communication Host names Labels separated by dots, e.g., www.example.org Final label is top-level domain Generic: .com, .org, etc.

Country-code: .us, .il, etc. Domains are divided into second-level domains, which can be further divided into subdomains, etc. E.g., in www.example.com, example is a second-level domain A host name plus domain name information is called the fully qualified domain name of the computer Above, www is the host name, www.example.com is the FQDN nslookup program provides command-line access to DNS (on most systems) looking up a host name given an IP address is known as a reverse lookup Recall that single host may have multiple IP addresses. Address returned is the canonical IP address specified in the DNS system ipconfig (on windows) can be used to find the IP address (addresses) of your machine ipconfig /displaydns displays the contents of the DNS Resolver Cache (ipconfig /flushdns to flush it) ANALOGY TO TELEPHONE NETWORK IP ~ the telephone network TCP ~ calling someone who answers, having a conversation, and hanging up UDP ~ calling someone and leaving a message DNS ~ directory assistance HIGHER-LEVEL PROTOCOLS Many protocols build on TCP Telephone analogy: TCP specifies how we initiate and terminate the phone call, but some other protocol specifies how we carry on the actual conversation Some examples: SMTP (email) (25) FTP (file transfer) (21) HTTP (transfer of Web documents) (80) WORLD WIDE WEB Originally, one of several systems for organizing Internet-based information

Competitors: WAIS, Gopher, ARCHIE Distinctive feature of Web: support for hypertext (text containing links) Communication via Hypertext Transport Protocol (HTTP) Document representation using Hypertext Markup Language (HTML) The Web is the collection of machines (Web servers) on the Internet that provide information, particularly HTML documents, via HTTP. Machines that access information on the Web are known as Web clients. A Web browser is software used by an end user to access the Web. HYPERTEXT TRANSPORT PROTOCOL (HTTP) HTTP is based on the request-response communication model: Client sends a request Server sends a response HTTP is a stateless protocol: The protocol does not require the server to remember anything about the client between requests. Normally implemented over a TCP connection (80 is standard port number for HTTP) Typical browser-server interaction: User enters Web address in browser Browser uses DNS to locate IP address Browser opens TCP connection to server Browser sends HTTP request over connection Server sends HTTP response to browser over connection Browser displays body of response in the client area of the browser window The information transmitted using HTTP is often entirely text Can use the Internets Telnet protocol to simulate browser request and view server response

HTTP
Connect

Send Request Receive Response

{ {

$ telnet www.example.org 80 Trying 192.0.34.166... Connected to www.example.com (192.0.34.166). Escape character is ^]. GET / HTTP/1.1 Host: www.example.org HTTP/1.1 200 OK Date: Thu, 09 Oct 2003 20:30:49 GMT

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

28

HTTP REQUEST Structure of the request: o o o o Start line o Example: GET / HTTP/1.1 Three space-separated parts: o HTTP request method o Request-URI (Uniform Resource Identifier) o HTTP version Request-URI Uniform Resource Identifier (URI) start line header field(s) blank line optional body

o Syntax: scheme : scheme-depend-part Ex: In http://www.example.com/ the scheme is http

o Request-URI is the portion of the requested URI that follows the host name (which is supplied by the required Host header field) Ex: / is Request-URI portion of http://www.example.com/

URIs are of two types: o Uniform Resource Name (URN) Can be used to identify resources with unique names, such as books (which have unique ISBNs) Scheme is urn

o Uniform Resource Locator (URL) Specifies location at which a resource can be found

In addition to http, some other URL schemes are https, ftp, mailto, and file HTTP request method Common request methods: GET Used if link is clicked or address typed in browser No body in request with GET method POST Used when submit button is clicked on a form Form information contained in body of request HEAD Requests that only header fields (no body) be returned in the response HEADER FIELD(S) Header field structure: field name : field value Syntax

10

Field name is not case sensitive Field value may continue on multiple lines by starting continuation lines with white space Field values may contain MIME types, quality values, and wildcard characters (*s) Multipurpose Internet Mail Extensions (MIME) Convention for specifying content type of a message In HTTP, typically used to specify content type of the body of the response MIME content type syntax: top-level type / subtype Examples: text/html, image/jpeg HTTP Quality Values and Wildcards Example header field with quality values: accept: text/xml,text/html;q=0.9, text/plain;q=0.8, image/jpeg, image/gif;q=0.2,*/*;q=0.1 Quality value applies to all preceding items Higher the value, higher the preference Note use of wildcards to specify quality 0.1 for any MIME type not specified earlier Common header fields: Host: host name from URL (required) User-Agent: type of browser sending request Accept: MIME types of acceptable documents Connection: value close tells server to close connection after single request/response Content-Type: MIME type of (POST) body, normally application/x-www-formurlencoded Content-Length: bytes in body Referer: URL of document containing link that supplied URI for this HTTP request HTTP RESPONSE

11

Structure of the response: status line header field(s) blank line optional body Status line Example: HTTP/1.1 200 OK Three space-separated parts: HTTP version status code reason phrase (intended for human use) Status code Three-digit number First digit is class of the status code: 1=Informational 2=Success 3=Redirection (alternate URL is supplied) 4=Client Error 5=Server Error Other two digits provide additional information Common header fields: Connection, Content-Type, Content-Length Date: date and time at which response was generated (required) Location: alternate URI if status is redirection Last-Modified: date and time the requested resource was last modified on the server Expires: date and time after which the clients copy of the resource will be out-ofdate

12

ETag: a unique identifier for this version of the requested resource (changes if resource changes) CLIENT CACHING A cache is a local copy of information obtained from some other source Most web browsers use cache to store requested resources so that subsequent requests to the same resource will not necessarily require an HTTP request/response o Ex: icon appearing multiple times in a Web page

Client

Client Caching
1. HTTP request for image 2. HTTP response containing image

Server

Browser

Web Server

3. Store image

Cache

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

50

Client

Client Caching

Server

Browser I need that image again

Web Server

Cache

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

51

13

Client

Client Caching
This HTTP request for image

Server

Browser I need that image again

Web Server

Cache

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

52

Client

Client Caching

Server

Browser I need that image again

Web Server

Cache

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

53

Cache advantages (Much) faster than HTTP request/response Less network traffic Less load on server Cache disadvantage Cached copy of resource may be invalid (inconsistent with remote version) Validating cached resource:

14

Send HTTP HEAD request and check Last-Modified or ETag header in response Compare current date/time with Expires header sent in response containing resource If no Expires header was sent, use heuristic algorithm to estimate value for Expires Ex: Expires = 0.01 * (Date Last-Modified) + Date CHARACTER SETS Every document is represented by a string of integer values (code points) The mapping from code points to characters is defined by a character set Some header fields have character set values: Accept-Charset: request header listing character sets that the client can recognize Ex: accept-charset: ISO-8859-1,utf-8;q=0.7,*;q=0.5 Content-Type: can include character set used to represent the body of the HTTP message Ex: Content-Type: text/html; charset=UTF-8 Technically, many character sets are actually character encodings An encoding represents code points using variable-length byte strings Most common examples are Unicode-based encodings UTF-8 and UTF-16 IANA maintains complete list of Internet-recognized character sets/encodings Typical US PC produces ASCII documents US-ASCII character set can be used for such documents, but is not recommended UTF-8 and ISO-8859-1 are supersets of US-ASCII and provide international compatibility UTF-8 can represent all ASCII characters using a single byte each and arbitrary Unicode characters using up to 4 bytes each ISO-8859-1 is 1-byte code that has many characters common in Western European languages, such as WEB CLIENTS Many possible web clients: Text-only browser (lynx)

15

Mobile phones Robots (software-only clients, e.g., search engine crawlers) etc. Web Browsers First graphical browser running on general-purpose platforms: Mosaic (1993)

Primary tasks: Convert web addresses (URLs) to HTTP requests Communicate with web servers via HTTP Render (appropriately display) documents returned by a server HTTP URLs

16

HTTP URLs
http://www.example.org:56789/a/b/c.txt?t=win&s=chess#para5 host (FQDN) authority port path query fragment

Request-URI

Browser uses authority to connect via TCP Request-URI included in start line (/ used for path if none supplied) Fragment identifier not sent to server (used to scroll browser client area)
Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides 63

WEB BROWSERS o Standard features o Save web page to disk o Find string in page o Fill forms automatically (passwords, CC numbers, ) o Set preferences (language, character set, cache and HTTP parameters) o Modify display style (e.g., increase font sizes) o Display raw HTML and HTTP header info (e.g., Last-Modified) o Choose browser themes (skins) o View history of web addresses visited o Bookmark favorite pages for easy return o Additional functionality: o Execution of scripts (e.g., drop-down menus) o Event handling (e.g., mouse clicks) o GUI for controls (e.g., buttons)

17

o Secure communication with servers o Display of non-HTML documents (e.g., PDF) via plug-ins WEB SERVERS Basic functionality: Receive HTTP request via TCP Map Host header to specific virtual host (one of many host names sharing an IP address) Map Request-URI to specific resource associated with the virtual host File: Return file in HTTP response Program: Run program and return output in HTTP response Map type of resource to appropriate MIME type and use to set Content-Type header in HTTP response Log information about the request and response httpd: UIUC, primary Web server c. 1995 Apache: A patchy version of httpd, now the most popular server (esp. on Linux platforms) IIS: Microsoft Internet Information Server Tomcat: Java-based Provides container (Catalina) for running Java servlets (HTML-generating programs) as back-end to Apache or IIS Can run stand-alone using Coyote HTTP front-end Some Coyote communication parameters: Allowed/blocked IP addresses Max. simultaneous active TCP connections Max. queued TCP connection requests Keep-alive time for inactive TCP connections Modify parameters to tune server performance Some Catalina container parameters:

18

Virtual host names and associated ports Logging preferences Mapping from Request-URIs to server resources Password protection of resources Use of server-side caching TOMCAT WEB SERVER HTML-based server administration Browse to http://localhost:8080 and click on Server Administration link localhost is a special host name that means this machine

19

Some Connector fields: Port Number: port owned by this connector Max Threads: max connections processed simultaneously Connection Timeout: keep-alive time

Each Host is a virtual host (can have multiple per Connector) Some fields: Host: localhost or a fully qualified domain name Application Base: directory (may be path relative to JWSDP installation directory) containing resources associated with this Host

20

Context provides mapping from Request-URI path to a web application Document Base field is directory (possibly relative to Application Base) that contains resources for this web application For this example, browsing to http://localhost:8080/ returns resource from c:\jwsdp-1.3\webapps\ROOT Returns index.html (standard welcome file) Access log records HTTP requests Parameters set using AccessLogValve Default location: logs/access_log.* under JWSDP installation directory Example common log format entry (one line): www.example.org - admin [20/Jul/2005:08:03:22 -0500] "GET /admin/frameset.jsp HTTP/1.1" 200 920 Other logs provided by default in JWSDP: Message log messages sent to log service by web applications or Tomcat itself logs/jwsdp_log.*: default message log logs/localhost_admin_log.*: message log for web apps within /admin context

21

System.out and System.err output (exception traces often found here): logs/launcher.server.log Access control: Password protection (e.g., admin pages) Users and roles defined in conf/tomcat-users.xml Deny access to machines Useful for denying access to certain users by denying access from the machines they use List of denied machines maintained in RemoteHostValve (deny by host name) or RemoteAddressValve (deny by IP address) SECURE SERVERS Since HTTP messages typically travel over a public network, private information (such as credit card numbers) should be encrypted to prevent eavesdropping

https URL scheme tells browser to use encryption Common encryption standards: Secure Socket Layer (SSL) Transport Layer Security (TLS)

Id like to talk securely to you (over port 443) HTTP Requests Heres my certificate and encryption data

Secure Servers
HTTP Requests

Heres an encrypted HTTP request

Heres an encrypted HTTP request HTTP Responses Heres an encrypted HTTP response

HTTP Responses

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

83

22

Secure Servers Man-in-the-Middle Attack

Whats IP address for 100.1.1.1 www.example.org?

Browser

Real www.example.org

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

84

Secure Servers Preventing Man-in-the-Middle

Whats IP address for 100.1.1.1 www.example.org?

Browser

Real www.example.org

Guy-Vincent Jourdan :: CSI 3140 :: based on Jeffrey C. Jacksons slides

85