Internet Protocols - A Primer

The internet is a network of multiple computers that communicate with each other. This network can be used for any number of purposes, from the World Wide Web and email, to authenticating software and games online. The languages used for this communication are called protocols. Each different use generally has a different protocol, and there are many, many of these. Some are throwbacks to the early days of computers, when developers created protocols for fun, and others are created for very specific purposes. There are a number of widely used protocols, and we’ll look at the four most likely to affect you.

Protocols are basically made up of a set number of commands and responses. They work in a simple request-response model: your computer sends a request to the server using one of the protocol’s commands, and the server returns a response. The exact format of the request and the nature of the response is defined by the protocol.

HTTP

HyperText Transfer Protocol is, unsurprisingly, the protocol used to transfer hypertext documents. More generally, it is the protocol used by the World Wide Web to deliver web pages. You ask your browser to use this protocol every time you request a website with the http:// prefix.

HTTP is a stateless protocol, which means that every request is handled on its own, and, once handled, all information on the server regarding this request is thrown away. A request cannot reference a previous request for any reason.

HTTP requests are simple requests for files. Everything else that can be done on the internet is managed by clever management of this one fact. Almost every HTTP server (or web server) today has extensions added to it that allow for the requested files to be dynamically generated as they are requested, and it is these extensions (eg: ASP.NET or PHP) that provide most of the power to websites. Without them, the internet would simply be a way to browse static hypertext documents. This is what it was designed for, and this is still how it works.

HTTP does allow for information to be included with the request, and this allows forms of data to be uploaded and even caters for file uploads. There are severe limits to the size of these requests, however, and for good reason. The HTTP protocol, and the servers that run it, are designed to serve large numbers of small files in response to a large number of quick requests. In order to do this efficiently, web servers use techniques that work very well for small requests, but are very inefficient for large requests. So, large requests will place extreme pressure on the resources available to the server. The default limit is traditionally 2Mb, but some hosts have increased this to 5Mb or even 10Mb. If you need to upload a file larger than that, you’ll need to use FTP.

FTP

File Transfer Protocol is used for (you guessed it) transferring files. It is specifically designed to transfer large numbers of files, and very large files, as efficiently as possible. You can use your browser to download files using the FTP protocol, but a much better way to do it is to use an FTP client. There are many out there, but a good open-source (and therefore free) client is FileZilla .

The FTP protocol is not stateless. Every request-response sentence is part of a conversation, which allows FTP to communicate large amounts of data more effectively, and be more secure, than HTTP. It also allows for extremely useful features such as resumed downloads, so if your internet connection drops at the end of downloading a very large file (or in the middle of downloading a very large list of small files), you can simply resume the download from where it left off, rather than starting again from scratch.

POP

Post Office Protocol is used to fetch your email from your email server. It is a simple protocol, which will connect you to your server, authenticate you, and provide you access to your email. You can retrieve a list of all the headers, download the actual email, and choose whether or not to leave a copy of those emails on the server, or to delete them once they’ve been downloaded. This is pretty much all there is to POP – actually sending emails uses a different protocol: SMTP.

POP is not the only way to access your email, though. Some servers still support IMAP (Internet Messaging Access Protocol), but your most likely alternative to POP access is webmail. Different webmail systems interact with your email server in different ways, and it is possible that the webmail interface may in fact use POP itself, but many interact directly with the server, and do not use a protocol at all.

SMTP

Simple Mail Transfer Protocol is used to send email. It will get the email from your computer to the email server of the recipient, so that they can fetch it when they want, using POP or webmail. As the name implies, this is a simple protocol, but it has been expanded over the years to provide some fairly complex capabilities, including the attachment of large files to email, and the ability to format the email in a number of ways, including HTML.

You should always remember that, at its core, email is designed to send a simple, textual, message from one recipient to another. All other embellishments are extensions to the original, and are only possible because they’re supported by the email clients (eg: Outlook or Pegasus). In particular, you should remember that many of the cool things you can do with email using Outlook will only work if the recipient is also using Outlook. This is fine when you know exactly who you’re sending the email to, but when you’re sending messages to a mailing list, you should be extremely careful not to include elements that will only work in Outlook.

For example, Gmail will very carefully check the HTML of any email it delivers before displaying it. It will strip out anything it doesn’t like, including background images. So, if you want to send an email from Outlook using stationary, any user on Gmail will receive the email, but without the stationary. If your cool email design requires displaying information in the background image to make sense, Gmail users won’t understand your mail at all.