Client-Server and Peer-to-Peer Architectures

In a network, a device requesting a service from a server.

A computer program or device providing a service to a client.

A network device connecting devices in a network.

A device that directs data packets on a network.

Data transmission only in one direction.

Network topology where devices connect to a central device.

Network topology where devices connect along a single cable.

Network topology where each device connects to every other.

A network topology where devices connect in a closed loop, with data flowing in one direction. Each device is connected to exactly two other devices.

A break in the ring disrupts the entire network.

A network topology that branches out from a central node (trunk). Each device connects to the trunk with a single path.

A network connecting personal devices like computers, smartphones, and tablets for low-level data transmission. Often a master device acts as gateway to a larger network.

A computer network covering a limited area, typically a home, school, office building, or campus.

A set of rules that dictate how data is formatted, ordered, transmitted, and received among network devices.

The part of a network that connects to end-users (clients and servers).

The part of a network that forms the backbone, connecting various edges and the interconnected routers forming the network of networks.

The time it takes for a packet to travel from source to destination. It's broken down into four components: processing delay, queuing delay, transmission delay, and propagation delay.

Time spent by a router processing a packet, including error checking and determining the output link. Typically very short, usually less than a millisecond.

Time spent by a packet waiting in a queue at a router's outgoing link. Depends on network congestion: higher congestion leads to longer queues.

A multiplexing technique where different signals are transmitted on separate frequency bands. Each call gets its own band, allowing for simultaneous transmission within that limited bandwidth.

A multiplexing technique that divides time into slots. Each call gets allocated specific slots, allowing for transmission only during those slots but at a higher bandwidth.

Time required to transmit a packet through a link. Calculated by dividing the packet's length (in bits) by the link's transmission rate (in bits per second).

Time taken for a packet to travel the physical distance of a link. Determined by dividing the link's length by the signal's propagation speed (approximately 2x10^8 meters per second).

A large, global Internet Service Provider (ISP) that acts as a backbone for the internet, connecting to other regional ISPs.

The rate at which bits are transmitted from sender to receiver. Can be measured instantaneously (at a specific point in time) or averaged over a longer period.

An Internet Service Provider (ISP) that operates within a specific geographic region and connects to other ISPs, including Tier-1 ISPs.

A physical location where multiple ISPs connect and exchange traffic directly, allowing for faster and cheaper data transfer between them.

Software designed to harm or disrupt a computer system. It can spread through various means, including email attachments, infected websites, and malicious downloads.

Google operates a vast network of data centers, interconnected by a private network, that bypasses higher tiers of the internet by peering directly with lower-tier ISPs.

An attack that aims to overwhelm a server or network resource with illegitimate traffic, making it unavailable to legitimate users.

IXPs provide a cost-effective and efficient way for ISPs to exchange traffic, reducing costs and improving performance for internet users.

Google connects with lower-tier ISPs directly or through IXPs, minimizing reliance on higher-tier ISPs, which improves speed and efficiency for their services.

A network architecture where clients request services from servers, which are always-on, have permanent IP addresses, and often reside in data centers. Clients are intermittently connected and may have dynamic IP addresses. Clients do not communicate directly with each other.

A network architecture where end systems directly communicate, acting as both clients and servers, requesting and providing services to each other. Peers are intermittently connected and may have dynamic IP addresses.

A program running within a host. Processes within the same host communicate using inter-process communication mechanisms, while processes in different hosts communicate by exchanging messages.

An endpoint for communication between processes. Processes can send and receive messages to/from their sockets.

A set of rules that defines the types, syntax, semantics, and rules for exchanging messages between applications. Examples include HTTP, SMTP.

Ensuring that data transmitted over a network arrives at its destination without any corruption or alteration.

The delay experienced by data packets travelling over a network. Some applications require low delays (e.g., real-time communication), while others can tolerate higher delays.

Cookies are small files stored on a user's computer, used by websites to maintain state between user interactions. This allows websites to remember preferences, track shopping carts, and maintain authorization.

A Web cache acts as a temporary storage for web content, reducing the need to fetch the same content repeatedly from the origin server. This can be on a client machine or a proxy server, speeding up web access and reducing server load.

Conditional GET is a technique used in web caching to prevent redundant data transfer. It allows the client to tell the server if it has a current copy of the resource, avoiding unnecessary downloads.

HTTP/1.1 enables sending multiple HTTP GET requests over a single TCP connection, allowing for streamlined data transfer. However, this can lead to 'head-of-line blocking' where the transmission of small objects is delayed due to larger objects preceding them.

HTTP/2 aims to reduce the delay in multi-object HTTP requests, enhancing web performance. It overcomes the limitations of HTTP/1.1 by enabling flexible order of object transmission, using priorities, and pushing unrequested objects to the client.