|
Creado por Julian Rottenberg
hace alrededor de 6 años
|
|
Pregunta | Respuesta |
Internet Checksum Example | - Note -> When adding numbers, a carryout from the most significant bit needs to be added to the result |
Internet Checksum Example (add two 16-bit integers) | |
Transport Control Protocol (TCP) (Point-to-point) | - One sender, one receiver |
Transport Control Protocol (TCP) (Reliable, in-order byte stream) | - no "message boundaries" |
Transport Control Protocol (TCP) (Pipelined) | - TCP congestion and flow control set window size |
Transport Control Protocol (TCP) (Full duplex data) | - Bi-directional data flow in same connection - MSS: maximum segment size |
Transport Control Protocol (TCP) (Connection-oriented) | - Handshaking (exchange of control msgs) init's sender, receiver state before data exchange |
Transport Control Protocol (TCP) (Flow controlled) | - Sender will not overwhelm receiver |
Transport Control Protocol (TCP) (Send & receive buffers) | |
TCP Segment Structure | |
TCP Sequence Numbers and ACKs (Seq. #'s) | - Byte stream "number" of first byte in segment's data |
TCP Sequence Number and ACKs (ACKs) | - Seq # of next byte expected from other side - Cumulative ACK |
TCP Sequence Numbers and Acks (How does receiver handle out-of-order segments?) | - TCP spec doesn't say, - up to implementor |
TCP Sequence Numbers and ACKs (Bild) | |
TCP Round Trip Time and Timeout (Recall) | - Reliable data transfer needs to handle timeouts |
TCP Round Trip Time and Timeout (How to set TCP timeout value) | - Longer than RTT -> But RTT varies - Too short: premature timeout -> Unnecessary retransmissions - Too long: slow reaction to segment loss |
TCP Round Trip Time and Timeout (How to estimate RTT?) | |
TCP Round Trip Time and Timeout | |
Example RTT Estimation | |
TCP Round Trip Time and Timeout | |
TCP Connection Establishment | - TCP connections can be established in active (connect) or passive mode (using listen/accept) - Note: The connection is established by the TCP-entities without further interaction with the application, i.e. there is no service primitive corresponding to T-CONNECT.Rsp |
TCP Connection Establishment (Active Mode) | - Requesting a TCP connection with a specified transport service user (identified via IP address and port number) |
TCP Connection Establishment (Passibe Mode) | - An application informs TCP, that it is ready to accept an incoming connection -> Can specify a specific socket, on which an incoming connection is expected, or -> all incoming connections will be accepted (unspecified passive open) -> Upon an incoming connection request, a new socket is created that will serve as connection endpoint |
Connection Identification in TCP (a TCP connection is setup) | - Between a single sender and a single receiver - More precisely, between application processes running on these systems - TCP can multiplex several such connections over the network layer, using the port numbers as Transport SAP identifiers |
Connection Identification in TCP | - (So |
TCP Connection Management (1) | |
TCP Connection Management (2) | |
TCP Connection Management (3) | |
A TCP Connection in all Three Phases (Connection Establishment) | - 3-Way-Handshake - Negotiation of window size and sequence numbers |
A TCP Connection in all Three Phases (Data transfer) | - Piggybacking of acknowledgements |
A TCP Connection in all Three Phases (Connection Release) | - Confirmed (!) - Avoids loss of data that has already been sent |
A TCP Connection in all Three Phases (Bild) | |
TCP Connection Management: State Transitions | |
TCP Connection Management: State Diagram | |
TCP Reliable Data Transfer | |
TCP Sender Events (Data received from application) | |
TCP Sender Events (Timeout) | - Retransmit segment that caused timeout - Restart timer |
TCP Sender Events (Ack received) | - If it acknowledges previously unacked segments -> Update what is known to be acked -> Start timer if there are outstanding segments |
TCP Sender (simplified) (Comments) | - SendBase-1: -> Last cumulatively ack'ed byte (so SendBase is next expected pkt) |
TCP Sender (simplified) (Example) | - SendBase = 72; y = 73, so the rcvr wants 73+; y > SendBase, so that new data is acked |
TCP Sender (simplified) (Bild) | |
TCP: Retransmission Scenarios (1) | |
TCP: Retransmission Scenarios (2) | |
TCP ACK Generation [RFC 1122, RFC 2581] | |
Fast Retransmit (Time-out period often relatively long) | - Long delay before resending lost packet |
Fast Retransmit (Detect lost segments via duplicate ACKs) | - Sender often sends many segments back-to-back - If segment is lost, there will likely be many duplicate ACKs |
Fast Retransmit | - If sender receives 3 ACKs for the same data, it supposes that segment after ACKed data was lost: -> Fast retransmit --> resend segment before timer expires |
Fast Retransmit Algorithm | |
Send and Receive Buffers in TCP (TCP maintains buffer) | - Sender, to service for error control - Receiver, to store packets not yet retrieved by application or received out of order -> Old TCP implementations used GoBack-N, and discarded out-of-order packets |
Send and Receive Buffers in TCP (Bild) | |
TCP Flow Control: Advertised Window | |
Nagle's Algorithm - Self-Clocking and Windows (TCP self-clocking) | - Arrival of an ACK is an indication that new data can be injected into the network (see also later) |
Nagle's Algorithm - Self-Clocking and Windows (What happens when an ACK for only small amount of data (e.g., 1 byte arrives)? (Send immediately?) | - Network will be burdened by small packets ("silly window syndrome") |
Nagle's Algorithm - Self-Clocking and Windows (Nagle's algorithm describes how much data TCP is allowed to send) | - When application produces data to send if both available data and advertised window >= MSS send a full segment, else if there is unpacked data in flight, buffer new data until MSS is full else send all the new data now |
Congestion Control in TCP (TCP's mechanism for congestion control) (Implicit feedback by dropped packets) | - Implicit feedback by dropped packets -> Whether the packets were dropped because queues were full or by a mechanism like RED is indistinguishable (and immaterial) to TCP - There are some proposals for explicit router feedback as well, but not part of original TCP - Assumption: Congestion is the only important source of packet drops! |
Congestion Control in TCP (TCP's mechanism for congestion control) (Window-based congestion control) | - I.e., TCP keeps track of how many bytes it is allowed to inject into the network as a window that grows and shrinks - Sender limits transmission (in addition to limit due to flow control): -> LastByteSent - LastByteAcked <= CongWin - Note: in the following discussion the flow control window will be ignored |
TCP ACK/Self-Clocking | - Suppose TCP has somehow determined a correct size of its congestion window -> Suppose also that the TCP source has injected this entire amount of data into the network but sill has more data to send - Thus: ACK not only serves as a confirmation, but also as a permit to inject a corresponding amount of data into the network -> ACK-clocking (self-clocking) behaviour of TCP |
TCP ACK/Self-Clocking (When to send more data?) | - Only acceptable when there is space in the network again - Space is available when packets leave the network - Sender can learn about packets leaving the network by receiving an acknowledgement! |
Good and Bad News (Good news: Ack arrival) | - Network could cope with the current offered load; it did not drop the packet - Let's be greedy and try to offer a bit more load - and see if it works => Increase congestion window |
Good and Bad News (Bad news: No ACK, timeout occurs) | - Packet has been dropped, network is overloaded - Put less load onto the network => Reduce congestion window |
Reduce Congestion Window by How Much? | |
Increase Congestion Window by How Much? | |
Additive Increase - Details | - Additive increase does not wait for a full RTT before it adds additional load - Instead, each arriving ACK is used to add a little more load (not a full packet) |
Additive Increase - Details (Specifically) | - Increment = MSS x (MSS / Congestion Window) - Congestion Window += Increment - Where MSS is the Maximum Segment Size, the size of the largest allow packet |
AIMD - Sawtooth Pattern of TCP's Offered Load (Summary) | - TCP uses an additive increase multiplicative decrease (AIMD) scheme |
AIMD - Sawtooth Pattern of TCP's Offered Load (Consequence) | - A TCP connection perpetually probes the network to check for additional bandwidth - Will repeatedly exceed it and fall back, owing to multiplicative decrease - Sawtooth pattern of TCP's congestion window/offered load -> Thi is simplified; we have to introduce one more mechanism! |
AIMD - Sawtooth Pattern of TCP's Offered Load (Bild) | |
Quickly Initialize a Connection: Slow Start | - Additive increase nice and well when operating close to network capacity - But takes a long time to converge to it for a new connection -> Starting at a congestion window of, say, 1 or 2 |
Quickly Initialize a Connection: Slow Start (Idea) | - Quickly ramp up the congestion window in such an initialization phase -> One option: double congestion window each RTT - Equivalently: add one packet per ACk - Instead of just adding a single packet per RTT |
Quick Initialize a Connection: Slow Start (Bild) | |
Leaving Slow Start | - When doubling congestion window, network capacity will eventually be exceeded - Packet loss and timeout will result - Congestion window is halved and TCP switches to "normal", linear increase of congestion window - The "congestion avoidance" phase |
Remaining Problem: Packet Bursts | |
Solution: Use Slow Start Here As Well | |
TCP Congestion Window Dynamics | |
Summary: TCP Congestion Control | - When CongWin is below Threshold, sender in slow-start phase, window grows exponentially - When CongWin is above Threshold, sender is in congestion-avoidance phase, window grows lineraly - When a triple duplicate ACK occurs, Threshold set to CongWin/2 and CongWin set to Threshold - When timeout occurs, Threshold set to CongWin/2 and CongWin is set to 1 MSS |
Summary: TCP Sender Congestion Control | |
Another Summary: TCP Congestion Control | - This description still glosses over some (minor) details, but captures the essence - Main source of complications: Stupidity of the network |
Another Summary: TCP Congestion Control (Extensions to TCP) | - Fast retransmit, fast recovery -> Take corrective actions without having to wait for a timeout -> Necessary for large delay*data rate networks |
Another Summary: TCP Congestion Control (Different TCP versions) | - TCP Tahoe, TCP Reno, TCP Vegas -> Main difference is the congestion control - Correct interoperation is a tricky question (e.g. fairness) - Complicated dynamics |
Short Advertisement For Those Who Want More On This... | |
Example: Evaluation of TCP Congestion Control | |
TCP Throughput (What's the average throughout of TCP as a function of window size and RTT?) | - For the sake of simplicity, let us ignore slow start - When W be the window size when loss occurs. - When window is W, throughput is W/RTT - Just after loss, window drops to W/2, throughput to W/2RTT - Average throughput: .75 W/RTT |
TCP Fairness (Fairness goal) | - If K TCP sessions share same bottleneck link of bandwidth R, each should have average rate of R/K |
TCP Fairness (Bild) | |
Why is TCP fair? (Two competing sessions) | - Additive increase gives slope of 1, as throughout increases - Multiplicative decreases throughput proportionally |
Why is TCP fair? (Bild) | |
Fairness (Fairness and UDP) | - Multimedia apps often do not use TCP -> Do not want rate throttled by congestion control - Instead, use UDP -> Pump audio/video at constant rate, tolerate packet loss - Research area: TCP friendly |
Fairness (Fairness and parallel TCP connections) | - Nothing prevents applications from opening parallel connections between 2 hosts - Web browsers do this - Example: link of rate R supporting 9 connections -> New application asks for 1 TCP, get rate R/10 -> New application asks for 9 TCPs, gets R/2! |
Delay Modeling (How long does it take to receive an object from a Web server after sending a request?) | - Ignoring congestion delay is influenced by: -> TCP connection establishment -> Data transmission delay -> Slow start |
Delay Modeling (Notation, assumptions) | - Assume one link between client and server of rate R - S: MSS (max. segement size, bits) - O: object size (bits) - no retransmissions (no loss, no corruption) |
Delay Modeling (Window size) | - First assume: fix congestion window, W segments - Then dynamic window, modelling slow start |
Fixed Congestion Window (1) | |
Fixed Congestion Window (2) | |
TCP Delay Modeling: Slow Start (1) | |
TCP Delay Modeling: Slow Start (Delay components) | - 2 RTT for connection estab. and request - O/R to transmit object - time server idles due to slow start - Server idles: -> P = min{K-1, Q} times |
TCP Delay Modeling: Slow Start (Example) | - O/S = 15 segments - K = 4 windows - Q = 2 - P = min{K-1, Q} = 2 - Server idles P=2 times |
TCP Delay Modeling: Slow Start (2) | |
TCP Delay Modeling (3) | |
TCP Delay Modeling (4) | |
Case Study: HTTP Modeling | |
HTTP Response Time (in seconds) (1) | |
HTTP Response Time (in seconds) (2) | |
Chapter Summary (Principles behind transport layer services) | - Addressing, multiplexing, demultiplexing - Connection control - Flow control - Congestion control |
Chapter Summary (Instantiation and implementation of the Internet) | - UDP - TCP |
Chapter Summary | - As we have seen, in TCP three important protocol functions are implemented "altogether" in one sliding window protocol |
Chapter Summary (Error control) | - by sequence numbers, ACKs & retransmissions |
Chapter Summary (Flow control) | - by looking at acknowledgements and permits (& seqnums) |
Chapter Summary (Congestion control) | - by further slowing down the sender if packets or ACKs get lost (assumption: packets mainly get lost because of congestion!) |
Some Network Applications | |
Creating a Network Application (Write programs that) | - Run on different end systems and -Communicate over a network - E.g., Web: -> Web server software communicates with browser software |
Creating a Network Application (No software written for devices in network core) | - Network core devices do not function at app layer - This design allows for rapid app development |
Creating a Network Application | |
Principles of Network Applications: Architectures (Principle alternatives) | - Client-server - Peer-to-peer (P2P - Hybrid of client-server and P2P |
Client-Server Architecture (Server) | - always-on host - permanent IP address - server farms for scaling |
Client-Server Architecture (Clients) | - communicate with server - may be intermittently connected - may have dynamic IP addresses - do not communicate directly with each other |
Pure P2P Architecture | - Not always on server - Arbitrary end systems directly communicate - Peers are intermittently connected and change IP address - Example: Gnutella => Highly scalable => But difficult to manage |
Pure P2p Architecture (Bild) | |
Hybrid of Client-Server and P2p ((Original) Napster) | - File transfer P2p - File search ceontralized: -> Peers register content at central server -> Peers query same central server to locate content |
Hybrid of Client-Server and P2p (Instant messaging) | - Chatting between two users is P2P - Presence detection/location centralized: -> User registers its IP address with central server when it comes online -> User contacts central server to find IP addresses of buddies |
Processes Communicating | - Within same host, two processes communicate using inter-process communication (defined by OS) - Processes in different hosts communicate by exchanging messages |
Processes Communicating (Process) | - program running within a host |
Processes Communicating (Client process) | - Process that initiates communication |
Processes Communicating (Server process) | - Process that waits to be contacted |
Processes Communicating (Note) | - Applications with P2P architectures have client processes & sever processes |
Sockets | - Process sends/receives messages to/from its socket |
Sockets (Socket analogous to door) | - Sending process shoves messages outdoor - Sending process relies on transport infrastructure on other side of door which brings messages to socket at receiving process |
Sockets (API) | - Application programming interface (API) -> (1) choice of transport protocol -> (2) ability to fix a few parameters |
Sockets (Bild) | |
Addressing Processes | - For a process to receive messages, it must have an identifier - A host has a unique32-bit IP address |
Addressing Processes (Does the IP address of the host on which the process runs suffice for identifying the process?) | - No, many processes can be running on same host |
Addressing Processes (Identifier) | - Identifier includes both the IP address and port numbers associated with the process on the host |
Addressing Processes (Example port numbers) | - Example port numbers: -> HTTP server: 80 -> Mail server: 25 |
Issues Defined by an Application-Layer Protocol (Types) | - Types of messages exchanged, e.g. request & response messages |
Issues Defined by an Application-Layer Protocol (Syntax of message types) | - Syntax of message types: what fields in messages & how fields are delineated |
Issues Defined by an Application-Layer Protocol (Semantics of the field) | - Semantics of the fields, ie, meaning of information in fields |
Issues Defined by an Application-Layer Protocol (Rules) | - Rules for when and how processes send & respond to messages |
Issues Defined by an Application-Layer Protocol (Open vs Proprietary protocols) (Public-domain protocols) | - open specification available to everyone - allows for interoperability - most protocols commonly used in the Internet are defined in RFCs - e.g. HTTP, FTP, SMTP |
Issues Defined by an Application-Layer Protocol (Open vs Proprietary protocols) (Proprietary protocols) | - defined by a vendor - specification often not publicly available - e.g., KaZaA |
What Transport Service does an Application Need? (Data loss) | - Some apps (e.g., audio) can tolerate some loss - Other apps (e.g., file transfer, telnet) require 100% reliable data transfer |
What Transport Service does an Application Need? (Timing) | - Some apps (e.g., Internet telephony, interactive games) require low delay to be "effective" |
What Transport Service does an Application Need? (Bandwidth) | - Some apps (e.g., multimedia) require minimum amount of bandwidth to be "effective" - Other apps ("elastic apps") make use of whatever bandwidth they get |
Transport Service Requirements of Common Applications | |
Internet Transport Protocols Services (TCP Service) (Connection-oriented) | - Setup required between client and server processes |
Internet Transport Protocols Services (TCP Service) (Reliable transport) | - between sending and receiving process |
Internet Transport Protocols Services (TCP Service) (Flow control) | - Sender won't overwhelm receiver |
Internet Transport Protocols Services (TCP Service) (Congestion control) | - Throttle sender when network overloaded |
Internet Transport Protocols Services (TCP Service) (Does not provide) | - Timing, minimum bandwidth guarantees |
Internet Transport Protocols Services (UDP Service) (Unreliable data transfer) | - Between sending and receiving process |
Internet Transport Protocols Services (UDP Service) (Does not provide) | - Connection setup, - Reliability, - Flow & congestion control, - Timing, or bandwidth guarantee |
Internet Applications & Transport Protocols | |
Web and HTTP | |
HTTP Overview | - Web's application layer protocol - HTTP 1.0: RFC 1945 - HTTP 1.1: RFC 2068 |
HTTP Overview (Client/Server model) (Client) | - browser that requests, receives, "displays" Web objects |
HTTP Overview (Client/Server model) (Server) | - Web server sends object in response to requests |
HTTP Overview (Bild) | |
HTTP Overview (Uses TCP) | |
HTTP Overview (HTTP is "stateless") | - Server maintains no information about past client requests |
HTTP Overview (Aside) | - Protocols that maintain "state" are complex! -> Past history (state) must be maintained -> If server/client crashes, their views of "state" may be inconsistent, must be reconciled |
HTTP Connections (Nonpersistent HTTP) | - At most one object is sent over a TCP connection - HTTP/1.0 uses nonpersistent HTTP |
HTTP Connections (Persistent HTTP) | - Multiple objects can be sent over single TCP connection between client and server - HTTP/1.1. uses persistent connections in default mode |
Nonpersistent HTTP (1) | |
Nonpersistent HTTP (2) | |
Response Time Modeling (Definition of RTT) | - time to send a small packet to travel from client to server and back |
Response Time Modeling (Response time) | - one RTT to initiate TCP connection - one RTT of HTTP request and first few bytes of HTTP response to return - file transmission time => total = 2RTT + transmit time |
Response Timing Model (Bild) | |
Persistent HTTP (Nonpersistent HTTP issues) | - Requires 2 RTTs per object - OS must work and allocate host resources for each TCP connection - But browsers often open parallel TCP connections to fetch referenced objects |
Persistent HTTP (Persistent HTTP) | - Server leaves connection open after sending response - Subsequent HTTP messages between same client/server are sent over connection |
Persistent HTTP (Persistent without pipelining) | - Client issues new request only when previous response has been received - One RTT for each referenced object |
HTTP Request Message | - Two types of HTTP messages: request, response - HTTP request message -> ASCII (human-readable format) |
HTTP Request Message (Bild) | |
HTTP Request Message: General Format | |
Uploading Form Input (Post method) | - Web page often includes from input - Input is uploaded to server in entity body |
Uploading From Input (URL method) | - Uses GET method - Input is uploaded in URL field of request line: => www.somesite.com/animalsearch?monkeys&banana |
Method Types (HTTP/1.0) | - GET -POST -HEAD -> Asks server to leave requested object out of response |
Method Types (HTTP/1.1) | - GET, POST, HEAD - PUT -> Uploads file in entity body to path specified in URL field - DELETE -> Deletes file specified in the URL field |
HTTP Response Message | |
HTTP Response Status Codes | |
Trying Out HTTP (Client Side) for Yourself | |
User-Server State: Cookies (Four components) | - Many major Web sites use cookies - Four components: -> 1) Cookie header line in the HTTP response message -> 2) Cookie header line in HTTP request message -> 3) Cookie file kept on user's host and managed by user's browser -> 4) Back-end database at Web site |
User-Server State: Cookies (Example) | - Susan access Internet always from same PC - She visits a specific e-commerce site for first time - When initial HTTP requests arrive at site, site creates a unique ID and creates an entry in backend database for ID |
Cookies: Keeping "State" | |
Cookies (What cookies can bring) | - Authorization - Shopping carts - Recommendations - User session state (Web e-mail) |
Cookies (Aside) (Cookies and privacy) | - Cookies permit sites to learn a lot about you - You may supply name and e-mail to sites - Search engines use redirection & cookies to learn yet more - Advertising companies obtain info across sites |
Web Caches (Proxy Server) | |
More About Web Caching | - Cache acts as both client and server - Typically cache is installed by ISP (university, company, residential ISP) |
More About Web Caching (Why Web caching?) | - Reduce response time for client request - Reduce traffic on an institution's access link - Internet dense with caches enables "poor" content providers to effectively deliver content (but so does P2P file sharing) |
Caching Example (1) | |
Caching Example (2) | |
Caching Example (3) | |
Conditional GET (Goal) | - don't send object if cache has up-to-date cached version |
Conditional GET (Cache) | - specify date of cache copy in HTTP request: If-modified-since: <date> |
Conditional GET (Server) | - response contains no object if cached copy is up-to-date: HTTP/1.0 304 Not Modified |
Conditional GET (Bild) | |
Basic Web Server Tasks (1) | |
Basic Web Server Tasks (Prepare and accept requests) | |
Basic Web Server Tasks (Read and Process) | |
Basic Web Server tasks (Respond to Request) | |
Web Server Architectures (Four basic models) | - Process model - Thread model - In-kernel model - Event-driven model |
1. Process Model | |
1. Process Mode (Advantages) | - Synchronization when handling different requests inherent in process model - Protection between processes (one process crashes, other unaffected) |
¿Quieres crear tus propias Fichas gratiscon GoConqr? Más información.