The Internet Flashcards
Transmission Control Program
responsible for - fault tolerance: data transmitted between networks can be cached and re-sent if it fails the first time.
Internet Protocol (IP)
responsible for - end-to-end: there are no single central systems that can take the whole network down; each host can send/receive to others.
The internet can be loosely defined as …
“a series of internetworked systems”
a series of interconnected networks sharing data.
Packet-Switching
IP data is transmitted in a format known as a packet. A packet uses a data format we’ve seen before: metadata in headers, and a body with content. The headers are used to get the packet to its destination, while the body contains the information we’d like to transfer.
We refer to IP’s communication style as packet-switching. This is when a message is split up into separate “packets”, delivered to a destination, and reassembled as appropriate. Remember that IP’s primary responsibility as part of the Internet’s “double threat”, TCP/IP, is maintaining an end-to-end state. For this reason, IP isn’t concerned about whether packets are received by the client in sequential order, and may sometimes even lose packets altogether while in transit!
TCP/IP: Four Layers
Application
The Application Layer includes protocols related to user-facing data. Some of the protocols that’s we’ve discussed, like HTTP and FTP, operate in this layer. The TCP/IP model doesn’t care what type of application data is used; whatever is transmitted from the Transport Layer is considered Application Layer data.
Transport
The Transport Layer includes (you guessed it) transport protocols! We’ve already discussed the two best-known: TCP and UDP. This layer focuses on connectivity between clients and servers, and relies on the lower layers to establish network connectivity.
Internet
The Internet Layer is where IP lives. Data is processed in packets on this layer, and routing is primarily handled with IP addresses. The Internet layer focuses mostly on connecting separate networks together.
Link
The Link Layer includes our lower-level communication standards. Link Layer protocols aren’t concerned with the type of data being transported, but instead focus on getting data from one local network resource to another. We jump up to the Internet layer when dealing with resources on other networks.
Transport Protocols
Transport protocols act as our “delivery person”. IP is concerned with machine-to-machine communication, and HTTP is designed for application-to-application communication. Transport protocols bridge the gap and help our data cover the last mile between the network and software.
They use ports for addressing.
TCP
is a connection-oriented protocol, meaning it establishes a connection between two sockets. This connection acts as safeguard from other error-prone protocols underneath it, including IP and Ethernet. Pieces of data sent via TCP (referred to as segments) respect a strict order and verify when they have been received. This means that data can’t be “lost” across a TCP connection: if a segment is received out of order, the receiver will ask the transmitter to re-send the missing segment. This behavior makes TCP a reliable protocol.
User Datagram Protocol (UDP)
Scientists working with TCP found that they sometimes didn’t need all the order and reliability that TCP provided, and they were willing to trade that for raw speed. UDP is connection-less and provides no verification for whether data is received. Because of this, we refer to it as an unreliable protocol.
Hold on, though! By “unreliable”, we certainly don’t mean “useless”. UDP is used in lots of familiar places: real-time video sharing, voice-over-IP phone calls, and DNS all rely on UDP as their transport protocol of choice. These services prioritize speed over reliability, so it makes sense that they would forego TCP’s additional lag. If some data is lost along the way, that’s okay - for example, you might just see lower-quality video for a moment.
DNS
a distributed approach to providing easily-understood names for internetworked devices. Practically, it’s similar to a phone book: DNS allows us to look up a specific IP address by its domain.
domain
A website’s domain refers to the “friendly” name for the website’s host, or the server providing the site’s content. A domain differs from a URL in that the domain is only the server’s identifier
Each individual domain is represented by a set of name servers, which store information about the domain’s registered subdomains. Name servers will direct a client where they need to go - even if that’s another name server! We refer to this process of working out which name server we need as resolution. Eventually, we’ll reach a name server that can tell us the specific IP address for the full domain. We refer to this as the authoritative name server for our domain. It has the final say!
When trying to resolve a domain name, we start from the rightmost part (the TLD) and work our way to the left. We’ll stop once we’ve reached an authoritative server that can give a direct address for the domain we’re seeking.
Each name server maintains a zone file: a text file containing host names, IP addresses, and resource types.
top-level domain (TLD)
is the last part of the domain, appearing just before the URL begins pointing at application routes (usually indicated with /’s) or query parameters (indicated with a ? and &’s). The best known TLDs are .com, .net, and .org. TLDs are managed by special organizations that have demonstrated the ability to handle the immense workload involved, often known as domain registries.
second-level domain
You’ll often hear the TLD & second-level domain lumped together as “the domain”. This is the name most people associate with the website. Through domain registrars, consumers can purchase second-level domains. The registrar maintains a listing of each purchase.
subdomains
The best-known subdomain is www, though this is less-used on newer sites. Subdomains can usually be freely created by the consumer.
