1. Chapter 3 Application Layer Functionality and Protocols

2. Applications: Interface Between the Networks

3. 3 Application Layer: OSI and TCP/IP Models The Application layer, Layer seven, is the top layer of both the OSI and TCP/IP models. Provides the interface between the applications we use to communicate and the underlying network.

4. 4 Application layer protocols are used to exchange data between programs running on the source and destination hosts. There are many Application layer protocols and new protocols are always being developed. Email HTTP (www) HTTP

5. 5 Application Layer: OSI and TCP/IP Models Functionality of the TCP/IP application layer protocols fit roughly into the framework of the top three layers of the:  OSI model: Application, Presentation and Session layers. Most early TCP/IP application layer protocols were developed before the emergence of:  personal computers, graphical user interfaces and multimedia objects. These protocols implement very little of the functionality that is specified in the OSI model Presentation and Session layers.

6. 6 The Presentation Layer The Presentation layer has three primary functions:  Coding and conversion of Application layer data to ensure that data from the source device can be interpreted by destination device.  Compression of the data in a manner that can be decompressed by the destination device.  Encryption of the data for transmission and the decryption of data upon receipt by the destination. Compression and Coding formats:  Graphics Interchange Format (GIF)  Joint Photographic Experts Group (JPEG)  Tagged Image File Format (TIFF).

7. 7 The Session Layer Create and maintain dialogs between source and destination applications. Handles the exchange of information to:  initiate dialogs  keep them active  restart sessions that are disrupted or idle for a long period of time Most applications, like web browsers or e-mail clients, incorporate functionality of the OSI layers 5, 6 and 7.

8. 8 Application Layer: OSI and TCP/IP Models Common TCP/IP Protocols  Domain Name Service Protocol (DNS) is used to resolve Internet names to IP addresses.  Hypertext Transfer Protocol (HTTP) is used to transfer files that make up the Web pages of the World Wide Web.  Simple Mail Transfer Protocol (SMTP) is used for the transfer of mail messages and attachments.  Telnet, a terminal emulation protocol, is used to provide remote access to servers and networking devices.  File Transfer Protocol (FTP) is used for interactive file transfer between systems. Note: Usually a single server will function as a server for multiple applications

9. 9 RFCs: Request For Comments The protocols in the TCP/IP suite are generally defined by Requests for Comments (RFCs).  Maintained by IETF (Internet Engineering Task Force)  There are a few in there for fun - ftp://ftp.rfc-editor.org/in- notes/rfc1882.txt.

10. 10 Application Layer Software Network-Aware Applications  Applications are the software programs used by people to communicate over the network.  They implement the application layer protocols and are able to communicate directly with the lower layers of the protocol stack.  Email Clients  Web Browsers User applications Services System Operations Within the Application layer, there are two forms of software programs or processes that provide access to the network:  applications  services

11. 11 Application Layer Software Application layer Services  Other programs may need the assistance of Application layer services to use network resources such as:  File transfer  Network print spooling  These services are the programs that interface with the network and prepare the data for transfer. User applications Services System Operations

12. 12 Application Layer Software Application layer uses protocols that are implemented within applications and services.  Applications provide people a way to create messages.  Application layer services establish an interface to the network.  Protocols provide the rules and formats that govern how data is treated. Bottom line:  When discussing an application like "Telnet" we could be referring to the application, the service, or the protocol.

13. 13 Application Layer Protocol Functions Application layer protocols are used by both the source and destination devices during a communication session. The application layer protocols implemented on the source and destination host must match. Protocols: (This will become clearer later! Herding cats.)  Establish consistent rules for exchanging data.  Specify the structure and type of messages that are exchanged.  Types: Request, response, acknowledgement, error message, etc.  Defines the dialogues, ensuring with transmissions met by expected responses, and with the correct service invoked.

14. 14 Application Layer Protocol Functions Applications and services can use multiple protocols.  Encapsulate the protocol or encapsulated by this protocol  Invoke other protocols Using a web browser (HTTP):  May invoke:  DNS, ARP, ICMP  May use:  TCP, UDP, Ethernet, PPP  Uses  IP

15. 15 Client Server Model Client: the device requesting the information Server: the device responding to the request is called a server. The client begins the exchange by requesting data from the server. Server responds by sending one or more streams of data to the client. In addition to the actual data transfer, this exchange may also require control information, such as:  user authentication  the identification of a data file to be transferred

16. 16 Servers A server is usually a computer that contains information to be shared with many client systems.  Web server  Email server  File or database server  Applications server Some servers may require authentication of user account information and vary permissions. Example, if you request to upload data to the FTP server, you may have permission to write to your individual folder but not to read other files on the site.

17. 17 Servers The server runs a service, or process, sometimes called a server daemon. Daemons (like other services) typically run in the background and are not under an end user's direct control. Daemons are described as "listening" for a request from a client.  Programmed to respond whenever the server receives a request for the service provided by the daemon. When a daemon "hears" a request from a client:  It exchanges appropriate messages with the client, as required by its protocol,  Proceeds to send the requested data to the client in the proper format.

18. 18 Application Layer Services and Protocols Servers typically have multiple clients requesting information at the same time. For example, a Telnet server may have many clients requesting connections to it.  These individual client requests must be handled simultaneously and separately for the network to succeed.  The Application layer processes and services rely on support from lower layer functions to successfully manage the multiple conversations.

19. Application Layer Protocols

20. 20 HTTP (WWW) FTP SMTP (email) Telnet (file transfer) (remote login) DHCP (IP address resolution) DNS (file sharing) P2P (domain name resolution) (file sharing) SMB We will examine HTTP in detail.

21. 21 Data HTTP Header TCP Header IP Header Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer IP Packet Data Link Header Data Link Trailer Data HTTP Header TCP Header IP Header Data Link Header Data Link Trailer Reminder of encapsulation/decapsulation

22. 22 Focus on Application Header and/or Data We will examine how the application (header) and/or data communication with each other between the client and the server. “Later” we will look at what roles the other layers, protocols (TCP, IP, etc.) play. HTTP

23. 23 HTTP (HyperText Transfer Protocol) HTTP – The Web’s application layer protocol. RFC 1945 and RFC 2616 Implemented in:  Client program  Server program Current version: HTTP/1.1 Encapsulated in TCP (more later) HTTP Server HTTP Client HTTP

24. 24 HTTP (HyperText Transfer Protocol) Web page (also called a html document) Web page consists of objects  Objects (examples):  HTML file  JPEG image  GIF image  JAVA applet  Audio file Homepage Munz Udo: <META NAME="keywords" CONTENT="GWS-Lörrach, Gewerbeschule Lörrach, The base HTML file references other objects in the page.

25. 25 Web Browser - Client Browser – The user agent for the Web.  Displays requested Web page and provides navigational and configuration features. Browser and client may be used interchangeably in this discussion. HTTP has nothing to do with how a Web page is interpreted (displayed) by the client (browser). HTTP Client

26. 26 Web Server Web Server – Stores web objects, each addressable by a URL. Implement the server side of HTTP. Examples:  Apache  Microsoft Internet Information Server HTTP Server

27. 27 HTTP Request Message Request Message  Request line  Header lines ASCII Text Request line: Method field  GET, POST and HEAD  The great majority of Requests are GETs HTTP Client HTTP Server http://de.wikipedia.org/wiki/Hypertext_Transfer_Protocol

28. 28 HTTP Request Message Request Line GET - Browser/client is requesting an object /home/ - Browser is requesting this object in this directory (default is index.html) HTTP/1.1 - Browser implements the HTTP/1.1 (1.1 is backwards compatible with 1.0) Note: HTTP GET is also used by some P2P applications like Gnutella and Bittorrent. GET /munz-udo.de/ HTTP/1.1 Accept-Language: en-us User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1;.NET CLR 2.0.50727; Media Center PC 5.0;.NET CLR 3.0.04506; InfoPath.1) Host: www.munz-udo.de Connection: Keep-Alive

29. 29 HTTP Request Message Request Line GET: - Used by browser/client to request an object. POST: - Used when user has filled out a form and sending information to the server. (Forms do not have to use POST.) - Example: words in a search engine HEAD:- Similar to a GET, but the server will responds with a HTTP message but leaves out the requested object. PUT:- Used with Web publishing tools, upload objects. DELETE:- Used with Web publishing tools, delete objects. GET /~home/ HTTP/1.1 Accept-Language: en-us User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1;.NET CLR 2.0.50727; Media Center PC 5.0;.NET CLR 3.0.04506; InfoPath.1) Host: www.munz-udo.de Connection: Keep-Alive

30. 30 HTTP Request Message Header Lines Accept-Language: - User prefers this language of the object User-Agent: - The browser type making the request Host: - Host on which the object resides Connection: - Client/browser is telling the server to keep this TCP connection Open, known as a persistent connection. - We will talk about this later in TCP (transport layer) GET /~home/ HTTP/1.1 Accept-Language: en-us User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1;.NET CLR 2.0.50727; Media Center PC 5.0;.NET CLR 3.0.04506; InfoPath.1) Host: www.munz-udo.de Connection: Keep-Alive

31. 31 HTTP Response Message HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html HTTP Client HTTP Server

32. 32 HTTP Response Message HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html Response message:  Status line  Header lines  Entity body

33. 33 HTTP Response Message HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html Status Line HTTP/1.1 – Server is using HTTP/1.1 200 OK - Status code, request succeeded and information is returned in response

34. 34 HTTP Response Message HTTP/1.1 404 Status Codes 200 OK - Status code, request succeeded and information is returned in response. 301 Moved Permanently - Requested object has been permanently moved. 400 Bad Request - Generic error message, request not understood by server. 404 Not Found: -The requested document does not exist on server. 505 HTTP Version Not Supported - The requested HTTP protocol version not supported by server.

35. 35 HTTP Response Message HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html Header Lines Date: – Server is using HTTP/1.1 Server: - Status code, request succeeded and information is returned in response Last-Modified: – Date/time when object created or modified Content-Length: – Number of bytes in object being sent Connection: – Server going to close TCP connection after sending the requested object. Content-Type: – Object in entity body is HTML text

36. 36 HTTP Response Message HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html Entity Body <!DOCTYPE html PUBLIC etc. : – HTML text and other objects to be used by the browser/client

37. 37 HTTP Request and Response Messages GET /~rgraziani/ HTTP/1.1 Accept-Language: en-us User-Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0; SLCC1;.NET CLR 2.0.50727; Media Center PC 5.0;.NET CLR 3.0.04506; InfoPath.1) Host: www.munz-udo.de Connection: Keep-Alive HTTP Client HTTP Server HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 16:34:18 GMT Server: Apache/2.0.52 (Red Hat) Last-Modified: Thu, 15 Nov 2007 19:33:12 GMT Content-Length: 15137 Connection: close Content-Type: text/html HTTP

38. 38 User-Server Interaction: Cookies Web servers are considered stateless – they do not maintain state information, keep track of the user.  Higher performance – allowing the server to handle thousands of simultaneous TCP connections (later). Web servers use cookies to track users. Cookies defined in RFC 2109

39. 39 User-Server Interaction: Cookies Web server installs cookies on client when:  Accessed the web site for the first time (Web server does not know client by name.) and/or  User provides information to the web server. (Web server now knows client by name.) HTTP on Web server responds with a Set-cookie: header with an ID.  This ID is stored on the client’s computer. Each time client/browser accesses web site. The GET includes Cookie: or User_ID or similar with the ID. HTTP Client HTTP Server HTTP Requests (GET) now include ID HTTP: Response Set-cookie: ID HTTP Requests: GET (first time) Web server can now track clients activities on the web site.

40. 40 HTTP Request and Response Messages GET /jpeg/cap81/cam0.36705623.rgb888.enc HTTP/1.1 Cookie: SLSPOTNAME5=Cowells; SLSPOTNAME4=Waimea%20Bay; SLSPOTNAME3=Pipeline; SLSPOTNAME2=38th%20Ave%2E; SLSPOTNAME1=Cowells; SLSPOTID5=4189; SLSPOTID4=4755; SLSPOTID3=4750; SLSPOTID2=4191; SLSPOTID1=4189; OAX=R8bfwEbcU08ABCBu; USER_ID=5551212 ; HTTP Client HTTP Server HTTP/1.1 200 OK Date: Fri, 22 Feb 2008 19:00:15 GMT Server: Apache/1.3.34 (Unix) Last-Modified: Fri, 22 Feb 2008 18:51:47 GMT ETag: "760a31-18ce-47bf19c3" Accept-Ranges: bytes Content-Length: 6350 Keep-Alive: timeout=15, max=257 Connection: Keep-Alive Content-Type: text/plain HTTP: Cookie 5551212 included HTTP data customized for Udo Munz

41. 41 Web Caching Web cache or proxy server – Web cache satisfies HTTP requests on the behalf of the Origin Web server.  Own disk storage  Keeps copies of recently requested objects Typically installed at ISP or larger institutions. Advantages:  Reduces the response time for client requests, especially if there are any bottlenecks in the network.  Reduces traffic on institution’s access link to the ISP (Internet). Client Origin Server Orgin Server HTTP Request HTTP Response Web Cache or Proxy Server

42. 42 Web Caching 1. Client/browser sends HTTP Request to Web cache (Proxy server). 2. Web cache checks to see if it has a local copy of the object.  2a. Local copy: Web cache sends object to client’s browser.  2b. No Local copy: Web cache sends HTTP request to origin server. 3. Origin server sends object to Web cache. 4. Web cache stores a local copy of the object. 5. Web cache forwards copy of the object to the client browser. Note: TCP connections are also created between Client and Web Cache; Web cache and Origin server (later). Web Cache or Proxy Server Client Origin Server HTTP Request HTTP Response

43. 43 Web Caching (Extra) Problem – stale cache  Object may have been modified on Origin server since copy was cached by the Web cache. Solution – Conditional GET  Request method: GET  Includes header: If-Modified-Since: Web cache sends Conditional GET to Origin server to see if there is a newer version of the object.  No newer version: Send current local object  Newer version: Replaces current object and forwards newer version.

44. 44 Web Cache – Steps (Extra) 1. Client/browser sends HTTP Request to Web cache (Proxy server). 2. Web cache checks to see if it has a local copy of the object.  No local copy 3. Web cache sends HTTP request to origin server. 4. Origin Web server sends HTTP response with object requested 5. Web cache stores local copy of the object with last-modified date. 6. Web cache forwards object to client/browser. 1. Web cache receives another request for this HTTP object. 2. Web cache sends a Conditional GET to the Origin Web server, with If- modified-since: header. 3. Origin Web Server returns:  No change: HTTP Response message 304 Not Modified, no object.  Web cache sends local object.  Change: HTTP Response 200 OK, with object.  Web cache replace object and forwards update object.

45. 45 HTTPS HTTPS (Hypertext Transfer Protocol over Secure Socket Layer) is a URL scheme used to indicate a secure HTTP connection. HTTPS is not a separate protocol  combination of a normal HTTP interaction over an encrypted:  Secure Sockets Layer (SSL) or  Transport Layer Security (TLS) connection http://de.wikipedia.org/wiki/Hypertext_Transfer_Protocol_Secure http://www.softed.de/fachthema/https.aspx

46. 46 FTP (File Transfer Protocol) FTP was developed to allow for file transfers between a client and a server. Used to push and pull files from a server running the FTP daemon (FTPd). Uses get and put commands. RFC 959 FTP Client FTP Server

47. 47 Client initiates a TCP control connection with FTP server using port 21.  This connection remains open until the user quits the FTP application.  TCP port 21 connection includes:  Username and password is sent over TCP port 21.  Remote directory changes  This state information significantly reduces total number of sessions on server. For each file transferred, TCP opens and closes a TCP data connection on port 20. More later on TCP ports and connections. TCP data connection port 20 TCP control connection port 21 Username and password Change directory on Server Copy file from client to server – Connection Closed TCP data connection port 20 Copy file from server to client – Connection Closed TCP control connection port 21 Quit FTP Application – Connection Closed FTP (File Transfer Protocol)

48. 48 SMTP – Simple Mail Transfer Protocol Email – One of the killer applications of the Internet.

49. 49 Internet mail involves:  User agents  Allows users to read, reply, compose, forward, save, etc., mail messages  GUI user agents: Outlook, Eudora, Messenger  Text user agents: mail, pine, elm  Mail servers  Stores user mail boxes, communicates with local user agents and other mail servers.  SMTP  Principle application layer protocol for Internet mail  Sent over TCP  Mail access protocols: POP3, IMAP, HTTP SMTP POP3 IMAP User agent Mail server SMTP – Simple Mail Transfer Protocol

50. 50 SMTP  RFC 2821  Transfers messages from sender’s mail server to recipient’s mail server  Push protocol, not a pull protocol  Push (from client to server or server to server)  Pull (from server to client) Retrieving email  Historically, users would log into local mail server to read mail.  Since early 1990’s, clients use mail access protocols:  POP3  IMAP  HTTP SMTP POP3 IMAP User agent Mail server SMTP – Simple Mail Transfer Protocol

51. 51 SMTP – Simple Mail Transfer Protocol POP3 (Post Office Protocol)  RFC 1939  Limited functionality  Uses TCP port 110  Download-and-delete mode  Retrieves messages on server and store the locally  Delete messages on server  Download-and-keep mode  Does not delete messages on server when retrieved.  Problem  Difficult to access email from multiple computers – work and home.  Some email may have already been downloaded on another computer (work) – download-and-delete  To read email from another computer, must leave on server – download-and-keep  Does not provide means for user to create remote folders on mail server

52. 52 IMAP (Internet Message Access Protocol)  RFC 2060  Mail not downloaded, but kept on server  Received email is associated with user’s INBOX  Users can create and manage remote folders  Users can retrieve portions of the email:  Message header: Subject line and Sender Web-based email  Introduced with Hotmail in mid-1990’s  Communicates with remote mailbox using HTTP  HTTP is used to push (client to server) and pull the email (server to client) SMTP IMAP HTTP User agent Mail server SMTP – Simple Mail Transfer Protocol SMTP

53. 53 SMTP Mail software, processes used: MTA and MDA MUA (Mail User Agent) – Email client software. MTA (Mail Transfer Agent) – Software that governs transfer of email between mail servers.  Includes UNIX sendmail, Microsoft Exchange Server, Postfix, and Exim MDA (Mail Delivery Agent) – Software that governs transfer of email from mail servers to clients.  On Unix systems, procmail and maildrop are the most popular MDAs. MTA receives email from the client's MUA passes email to the MDA for final delivery uses SMTP to route email between servers

54. 54 Telnet Telnet provides a standard method of emulating text-based terminal devices over the data network. http://de.wikipedia.org/wiki/Telnet Server Telnet

55. 55 Telnet Allows a user to remotely access another device (host, router, switch). A connection using Telnet is called a Virtual Terminal (VTY) session, or connection. Telnet uses software to create a virtual device that provides the same features of a terminal session with access to the server command line interface (CLI). Telnet clients:  Putty  Teraterm  Hyperterm Server Telnet

56. 56 Telnet Telnet supports user authentication, but does not encrypt data. All data exchanged during a Telnet sessions is transported as plain text. Secure Shell (SSH) protocol offers an alternate and secure method for server access.  Stronger authentication  Encrypts data

57. 57 DHCP – Dynamic Host Configuration Protocol IP addresses and other information can be obtained:  Statically  Dynamically (DHCP)

58. 58 DHCP DHCP Information can include:  IP address  Subnet mask  Default gateway  Domain name  DNS Server DHCP servers can be:  Server on LAN  Router  Server at ISP

59. 59 DHCP We will discuss DHCP more when we discuss IPv4.

60. 60 DNS – Domain Name System DNS allows users (software) to use domain names instead of IP addresses

61. 61 Name Resolution Resolver DNS client programs used to look up DNS name information. Name Resolution The two types of queries that a DNS resolver (either a DNS client or another DNS server) can make to a DNS server are the following: Recursive queries Queries performed by Host to Local DNS Server Iterative queries Queries performed Local DNS server to other servers Need the IP address

62. 62 DNS Name Resolution User types http://www.example.com Step 1. The DNS resolver on the DNS client sends a recursive query to its configured Local DNS server. Requests IP address for "www.example.com". The DNS server for that client is responsible for resolving the name  Cannot refer the DNS client to another DNS server. 1

63. 63 Step 2. Local DNS Server forwards the query to a Root DNS server. Step 3. Root DNS server  Makes note of.com suffix  Returns a list of IP addresses for TLD (Top Level Domain Servers) responsible for.com. DNS Name Resolution 1 2 2 3

64. 64 DNS Name Resolution Root DNS Servers  There are 13 Root DNS servers (labeled A through M) TLD Servers  Responsible for domains such as.com, edu, org,.net,.uk, jp, fr  Network Solutions maintains TLD servers for.com  Educause maintains TLD servers for.edu  There are redundant servers throughout the world.

65. 65 Step 4. The local DNS server sends query for www.example.com to one of the TLD servers. Step 5. TLD Server  Makes note of example.com  Returns IP address for authoritative server example.com (such as dns.example.com server) DNS Name Resolution 4 4 5

66. 66 Step 6. Local DNS server sends query for www.example.com directly to DNS server for example.com Step 7. example.com DNS server responds with its IP address for www.example.com 6 6 DNS Name Resolution 7

67. 67 Step 8. Local DNS server sends the IP address of www.example.com to the DNS client. DNS Caching When a DNS server receives a DNS reply (mapping hostname to an IP address) it can cache the information in its local memory. DNS servers discard cached information after a period of time (usually 2 days) A local DNS server can cache TLD server addresses, bypassing the root DNS servers in the query chain. 8 7 DNS Name Resolution

68. 68 DNS Name Resolution In the worst cases, you'll get a dialog box that says the domain name doesn't exist - even though you know it does. This happens because the authoritative server is slow replying to the first, and your computer gets tired of waiting so it times-out (drops the connection) or the domain name does not exist. But if you try again, there's a good chance it will work, because the authoritative server has had enough time to reply, and your name server has stored the information in its cache.

69. 69 nslookup Displays default DNS server for your host Can be used to query a domain name and get the IP address

70. 70 DNS Name Resolution ipconfig /displaydns  After a certain amount of time, specified in the Time to Live (TTL) associated with the DNS resource record, the resolver discards the record from the cache.  ipconfig /flushdns – Manually deletes entries The default TTL for positive responses is 86,400 seconds (1 day). The default TTL for negative responses is 300 seconds.

71. 71

72. 72 SMB – Server Message Block Protocol The Server Message Block (SMB) is a client/server file sharing protocol. IBM developed Server Message Block (SMB) in the late 1980s to describe the structure of shared network resources, such as directories, files, printers, and serial ports.

73. 73 SMB Request-response protocol. Unlike FTP, clients establish a long term connection to servers. Client can access the resources on the server as if the resource is local to the client host. SMB is sent over TCP  Prior to Windows 2000 windows used a proprietary protocol (NETBIOS) to send SMB. Linux/UNIX have similar protocol: SAMBA

74. 74 SMB SMB messages can:  Start, authenticate, and terminate sessions  Control file and printer access  Allow an application to send or receive messages to or from another device

75. 75 Peer-to-Peer (P2P) Networking and Applications In addition to the client/server model for networking, there is also a peer-to-peer model. Two or more computers are connected via a network and can share resources (such as printers and files) without having a dedicated server. End devices (peers) can function as either a server or client.

76. 76 P2P File Sharing P2P (Peer-to-Peer) file sharing accounts for more traffic on the Internet than any other application (2004). Peers (hosts) act as both clients and servers. No centralized file server. HTTP GET and responses are commonly used.

77. 77 By Peter Svensson The Associated Press Oct. 19, 2007 “Peer-to-peer applications account for between 50 percent and 90 percent of overall Internet traffic, according to a survey this year by ipoque GmbH, a German vendor of traffic-management equipment.”

78. 78 Challenge with P2P – locating content across thousands or millions of peers. One solution – centralized directory  Approach done by Napster Problems (non-legal problems)  Single point of failure  Performance bottlenecks 1 – Inform and Update Centralized Directory Server P2P – Centralized Directory 1 – Inform and Update Peer 2 – Query for content 3 – File Transfer Napster

79. 79 1. Peer A starts P2P application 2. Informs centralized directory server of its:  IP address  Names of objects making available for sharing (MP3, videos, etc.) 3. Directory server collects information from each peer that becomes active.  Dynamic database  Maps IP addresses with object names 4. Peer A queries directory server for IP addresses of other peers for specific content  Directory Server returns IP addresses for those peers (Peer B) 5. Peer A establishes TCP connection and downloads file (i.e. HTTP GET) from other peer, Peer B. 6. Directory server removes Peer from database when Peer closes application or disconnects from Internet (periodic messages – pings – from server). 1 – Inform and Update Centralized Directory Server P2P – Centralized Directory 1 – Inform and Update Peer Peer B Peer Peer A 2 – Query for content 3 – File Transfer

80. 80 Gnutella – public domain file sharing application Fully distributed approach  No centralized server Gnutella peer maintains peering relationship (TCP connection – later) which a number of other peers (usually fewer than 10). P2P – Query Flooding Peer A Peer CPeer B Peer D Peer E Peer F Query Query hit File transfer Gnutella - Limewire

81. 81 Peer A searches for a file 1. Peer A sends query to all neighboring peers. 2. If neighboring peer does not have file, forwards query to all its neighboring peers 3. If any peer has the file it returns a query hit message. 4. Peer A selects a peer, Peer C, to retrieve file (HTTP GET) 5. A direct TCP connection is made with selected peer, Peer C. 6. HTTP response is used to send file. Query Flooding Non-scalable and causes a significant amount of traffic on Internet. Gnutella modified it to limited-scope flooding which limits how many peers away the query is sent to, usually 7 to 10. (similar to TTL – later). Peer A Peer CPeer B Peer D P2P – Query Flooding Peer E Peer F Query Query hit File transfer

82. 82 How a peer joins and departs Gnutella network 1. Finding peers:  Bootstrap program: Client maintains a list of peer IP addresses who are usually up  Contact Gnutella site that maintains a list 2. Client attempts to make contact with peers (TCP connection – later) 3. Client sends Gnutella ping message to peer.  Forwards Gnutella ping to other peers, who continue to forward ping until limited-scope is reached. 4. Each peer returns a Gnutella pong message including:  Its IP address  Number of files it is sharing  Total size of the files Peer A Peer CPeer B Peer D P2P – Query Flooding Peer E Peer F Query Query hit File transfer

83. 83 Kazaa combines ideas from Napster and Gnutella 2004 – Contributed to more traffic on Internet than any other application 2007 – Bittorrent became the leading application Proprietary technology P2P - Combination Kazaa

84. 84 Kazaa does not use a centralized server  Group leader peers (parent)  Higher bandwidth and Internet connectivity  Greater Gnutella responsibilites  Peers (child) – non-group leaders Child peer establishes TCP connection with a group leader Group leader:  maintains database directory of child peers including their IP addresses  maintain TCP connections with other group leaders Child peers query group leaders who forward the query to other group leaders Child peer selects peer for TCP connection and file transfer P2P - Combination Group Leader Query Query Reply File Transfer

85. Chapter 3 Application Layer Functionality and Protocols