In Networking Models, we learnt that the OSI model defines the physical layer as the bottommost layer of its model. This layer deals with the transfer of bits (binary data) converted into signals. The signal depends on the transformation medium used: electrical signals, light signals, radio waves etc.
Information is made of bits which can be defined as a pair of opposites: on or off, or, in reality, 1 and 0. A bits has two possible states, which is why we call a bit binary code.
Bits are like the atoms of information.
They are physically send by electricity, light and radio wave.
Now, to send a bit via electricity is pretty straightforward. However, to differenciate between one 0 and more than one 0 consecutive, we need to use a clock. The time per bits is agreed upon and that lets the machine counts how many zero or how many one between the agreed upon laps of time.
This is basically what is going on inside an Ethernet wire. An Ethernet wire is cheap but signal loss happens over just a few meters.
Something way faster is the light. We can send bits as light beam (light/no light binary) using fiber optic cable.
A fiber optic cable is a thread of glass engineered to reflect light. The bean of light bounces up and down the length of the cable until its destination. Depending on the bounce angle we can send multiple bits simultaneously, at the speed of light.
Wireless machine uses radio signal to send bits from one place to another. The machine translates the ones and zeroes into radio wave of different frequencies and the receiving machine reverse the process.
Bandwidth is the amount (maximum transmission capacity) of a device. It is measured by bit rate which is the number of bits that we can send over a given period of time, usually measured in seconds.
Increasing bandwidth doesn’t always result in a improvement of the network’s performance.
The bandwidth varies across the network. The core network is generally much buffier than the part of the “last-mile” (see below).
When a bandwidth changes from high to low at a certain point, we call this point a bandwidth bottleneck.
When dealing with a large amounts of data, low bandwidth is an issue. However most of the time latency is the biggest factor vis-a-vis of the performance of a networked application.
Latency is the time it takes for a bit to travel from the source to the requesting device. We can think of it as a “delay”.
Types of Delay
The total latency between two points is the sum of all the following delays:
- Propagation delay: the amount of time it takes for a message to travel from sender to receiver. Ratio between distance and speed.
Transmission delay: the amount of time it takes to add the data onto the link (between all the network devices, when using more than one network, which is typically the case on Internet). Can be checked with
- Processing delay: the amount of time necessary for our data to being processed in various ways.
- Queuing delay: the amout of time the data is queued or buffered when there is more data than the device can handle.
- Last-mile latency: generally the slowest link on the path. This is often the delays between the ISP’s network and the home network and is were most of the delays happen.
- Round-Trip Time (RTT): the amount of time for a signal to be sent + teh length of time for an acknowledgment/response to be received.
- Network Interface Card (NIC): a hardware component that understand the Ethernet protocol and translate (the job of an interface) it for the computer. On the other way, it translates the computer instructions into Ethernet protocol readable format.
- Network HUB: a hub is a hardware component that replicates the messages it receives and forwards it back to all of the devices connected to it, without distinction.
- Switches: similar to the hub, but unlike it, it reads the destination address (the MAC Addresses) in order to send the received messages only to the device for which it was intended.