Web Notes

1) The four sources of packet delay and the key factor(s) that affect each. (Section 1.4.1)

The total nodal delay = Processing delay at the node + queueing delay + transmission delay + propagation delay.

-Processing delay is caused by the processing that happens in the node (router, etc.) for things like checking bit errors and determining output link. This delay is minimal, as processing is usually less then 1 millisecond.

-Queueing delay occurs as packets stack up waiting to be transmitted. The amount of queuing delay depends on how busy the node is, with higher congestion causing greater queueing delay.

-Transmission delay is the delay caused by the limitations of the node to push packets into the network. It is expressed in terms of L/R, or packet length/bandwidth. Bigger packets and lower bandwidth would both result in higher transmission times.

-Propgation delay is the amount of time required to cover the physical distance over a transmission medium. The speed is typically near the speed of light for most transmission mediums, and is primarily affected by distance travelled. Length/speed = time of propagation delay.

2) The name and function of each layer in the IP stack. (Section 1.5.1)

·Application – Provide application-to-application services to users and programs by transmitting messages. Protocols at this layer include HTTP, SMTP, and FTP.

·Transport – Transport application layer messages between endpoints by sending segments, TCP and UDP are protocols at this layer. TCP offers, flow control and congestion control, and reliability, UDP does not.

·Network – Routes datagrams, which contain a destination IP address, from source to destination using the IP protocol. Routing protocols on networks decide how packets travel across the network.

·Link – Transmits frames from one the interface of one node to the interface of the next. Include protocols like Ethernet and WiFi for communicating between devices.

·Physical – Moves the physical bits “on the wire”. Different protocols for each transmission method, which could include twisted copper, coax, fiber, etc.

·Pneumonic “A tiny not long penis” + “My stinky dog farts biscuits”

3) Define IP address, socket, and protocol, and describe their role in process communications. (Sections 1.1.3, 1.5.2, 2.1.2)

An IP address is a 32 bit quantity which uniquely identifies a host on a network, and acts like a mailing address, to help locate the host. Protocols define format, order of messages sent and received among network entities, and actions taken on message transmission and receipt. Protocols are like rules to govern how information moves between hosts, and ensure pieces will know how to work together. A socket, also called the API, is the interface between process and the network. When a process is ready to communicate, it sends information out of the socket, and other processes receive the information through their socket.

4) The three primary protocols used for email. (Section 2.3)

SMTP – Uses TCP to for reliable email transfer from client to server (and server to client). Stores messages on mail server.

POP3 – Post Office Protocol, allows for simple user authorization and the ability to download messages from the mail server, as well as issue delete and other commands to the mail server. Won’t see email in other client once downloaded.

IMAP – Like POP3, but remembers state across sessions, stores all messages in server, and utilizes a file storage structure. What is most commonly used now to allow for multiple devices.

5) Dedicated and shared access networks as discussed. (Section 1.2.1)

Dedicated access networks provide a direct link between the ISP central office and the client. An example of dedicated access is DSL, or a direct fiber connection. A shared access method allows multiple connections over the same line, similar to a cable network. All packets in a shared access network travel on the same line, through each node on the link, to the central office.

6) FDM and TDM. (Section 1.3.2)

Frequency Division Multiplexing divides up the available bandwidth of a link into different frequencies, each of which provides a different connection across the single link. In Time Division multiplexing  

7) Throughput vs. bandwidth. (Section 1.4.4)

8) HTTP (non-persistent and persistent). (Section 2.2.2)

In HTTP, a TCP connection is established, a file request is made, HTTP messages are exchanged, and the TCP connection is closed. In non-persistent HTTP, only a single object is sent over the TCP connection, which mean the whole process in the first sentence must be repeated for every object that will be required to load a page. In persistent HTTP, the TCP connection remains open until all requests and transfers have been made. This greatly reduces overhead and improves loading time.

9) Web caching and its benefits. (Section 2.2.5)

A web cache sits close to a client, and serves client requests without involving the origin server. The cache stores recently requested objects on its own disks, and if a client requests something in the cache, it can serve it back quickly. If the object isn’t in the cache, the cache requests the object from the origin, stores it, and sends the requested copy to the client. This can speed response times, reduce link traffic, and deliver content better for small providers with no CDN.