First it is important to understand a few basic terms: A 'server' is a computer. It is unmanned and is located along with many others far, far away.
It's purpose is to receive, store and transfer data to whoever requests it. It serves up data - it is a server.
A 'router' is an unmanned piece of equipment that directs and routes electronic messages - much like a telephone exchange.
Servers and Routers often work together, so for the sake of this tutorial, they will all be classed as 'servers'.
Most home computers use a program ( an email client ) to access their emails. The most popular are Outlook, Outlook Express, Thunderbird, etc.
Some users access a client web site such as Yahoo, Hotmail or AOL to access their email client program - the result is the same.
Your emails will be handled by two specific servers, one for your Inbound mail and one for your Outbound mail.
The addresses and protocols of these servers are configured within your email client program.
Typically you can find this configuration data at Tools / Accounts / Mail / Properties / Servers ...
Sending an email: You fix the address of the recipient to your message in the form of their email address.
This determines which server and which account within that server will eventually hold that message.
The message is initially sent to your Outbound Email Server for journey preparation.
Receiving an email: You request from your Inbound Mail Server a download of all messages that are stored for you.
This may be done on demand or on a regular timed period determined by the preferences of your program.
If your email client program does not make a 'read' request your emails will remain unread at the Inbound Server.
Between your Outbound Server and my Inbound Server is the 'Internet', a vast array of interconnected servers of such mind-boggling technical complexity that we will avoid it's workings here. The term 'web' was not coined by accident. Imagine a massive spider's web, but in three dimensions - and a server at every junction. The idea is to route your message from A to B as efficiently as possible. At each stage the route may be re-calculated based on server or infrastructure workload. If a proposed route suddenly experiences a server that is showing overload or delays, the route will be amended to enable onward transfer. The result is that your email may get from A to B in one hop, or it may go through dozens of servers on different continents before arriving at its destination.
That was the easy part, and most emails manage to get from A to B without any problem.
In recent years 'spam' or unsolicited emails have clogged up the internet and have troubled every email account holder on the planet. It has been estimated that over 80% of all email traffic is 'spam'.
Many email servers and Internet Service Providers now use filter software to reduce the amount of unwanted email traffic. This works in one of two ways - it will either pass the message to its destination, but with a warning, or it will delete messages that it deems are illegal or 'spam'. The latter option may also delete perfectly good messages that the filter has incorrectly identified as 'spam'.
With the possibility of your cherished message travelling through countless intermediate servers there is no telling what the spam filtering policy might be on any one of them.
In the diagram above a message sent from 'Your Computer' to the 'Past Pages Computer' could possibly encounter spam detection software at (1) your ISP, (2) your Outbound Server, (3) Past Pages Inbound server, (4) Past Pages ISP, (5) Past Pages Computer. That's without the unknown servers lurking in the shadows if the 'Internet'.
The diagram below shows a simplified interpretation of server inter-connectivity.
Each server is shown with only 8 ports, but in reality they are countless and far too complicated for a simplified diagram.
The servers are not identical. They can be of any computer hardware and any be running any server software.
There is no formal structure to their inter-connectivity. The web is dynamic and ever changing. Access and availability is always changing.
Servers may come and servers may go. At any time some are being added, removed, taken off-line, updated, modified, hacked and attacked.
The internet was originally conceived as a military project by which, the integrity of reliable communications was maintained if some or many servers were destroyed.
The message would find a way through, based on each surviving server's ability to route the message via the best route available at the time.
The next element of distress is to learn that your carefully crafted email and any attachments will be chopped up before transmission.
All emails are sent via an Internet Protocol ( or IP for short), and a part of this specification limits the size of any one packet of information to about 1 Kilo-byte.
If your message and attachment is 50 KB in size it will be chopped up into about 50 packets, each one serial numbered, and sent on its way, wrapped in an overall IP message container - much like an envelope containing a letter.
It would be good at this point to imagine wanting to send a book through the post. First you rip out the pages, pack one page per envelope, number all the envelopes and post them.
One property of the internet and IP transmissions is time-sharing - a fair play system that will not let any one facility hog the show and dominate the communications highways. Your 50 packets of email will be automatically sprinkled onto the internet in between all the other packets in transit. No one has priority. Your outbound server may have several thousand emails at any one time to distribute, and it may deal with them by sending one packet from each message sequentially. This will result in many (or maybe all) of your packets taking different routes through different sets of servers. Packet No 22 may go via Liverpool and No 23 via Hong Kong! The route is of no consequence.
As an added complication, some equipment may only handle small packets ( maybe 0.5 KB), so the existing packets may be chopped even further.
At the receiver end, typically my Inbound Server, your message will be gathered in, all the packets arranged in order, and stuck back together as a unified message (or attachment). The system is clever enough to request lost or corrupted packets to be re-transmitted. If the assembly was successful my Inbound Server will inform my computer that I have 'Mail To Read'. On requesting the email to be downloaded from the server the data will be chopped up again for IP transmission to my computer.
If you've ever wondered why email transfers sometimes take seconds and at other times take hours, and sometimes never, well now you're a little wiser.
www.pastpages.co.uk Tony Nicholls 2008