There are two major types of network commonly used in our every day lives. We’re going to take a closer look.

WiFi: A LITTLE HISTORY

For most of us, WiFi has been around “forever”, but in truth, its roots can be traced back to 1971. The CSIRO are generally credited with inventing WiFi, and have patented a number of related technologies but WiFi, in fact, uses a large number of patented technologies from many developers.

In 1999, the WiFi Alliance was formed to establish and enforce standards for interoperability and backward compatibility, and to promote wireless local-area-network technology. The IEEE (Institute of Electrical and Electronics Engineers) establishes standards for many technologies, one of which is WiFi. Their 802.11 standards are the basis of WiFi protocols.

The Australian company, Netcomm experimented with “radio modems” inside cars racing in the V8 series many years ago. At the time it seemed pretty silly, but now it’s commonplace. Modern Formula 1 cars may have hundreds of sensors all transferring masses of data, and all of it is transmitted via WiFi.

Interestingly, the term “WiFi” is supposedly a pun related to HiFi, or high fidelity, or quality audio reproduction.

WHAT IS IT?

WiFi uses radio signals to transfer data between compatible devices. It operates in the 2.4GHz and/or 5GHz range. Each of these ranges are further divided into channels. The 2.4GHz range has 13 channels while the 5GHz band has many channels, but the number of 5GHz channels varies from country to country. While many devices will find a channel automatically, this process may not provide ideal results. To help avoid interference between devices, WiFi uses a low transmission power.

WHY ARE THERE TWO FREQUENCY BANDS?

As the popularity of WiFi has increased, the take-up of the technology has been dramatic. It is not uncommon to find six or more local networks competing for WiFi signal in your home. In the DIYODE office there are 12. Most modem/routers will use channel 6, as will likely your neighbours. This may lead to slow performance and/or intermittent loss of signal (i.e. a dropout). You can change the channel used by your modem/router fairly easily in most cases, but unless you know which channels are available you’re taking a chance. There is an excellent app for Android called “WiFi Analyser”. It will show you the signal strength of each channel and who is using what. Chances are you will find multiple devices trying to use the same channel.

As an alternative, if your equipment is suitable, you could use the 5GHz range which is currently less crowded. Testing in our office shows seven networks. High performance modems/routers use both the 2.4GHz and 5GHz bands.

THE EVOLUTION OF WIFI

The 802.11 stardard has undergone many versions with ever-increasing performance specifications.

The initial 802.11-1997 protocol allowed up to a whopping 2Mbps transfer rate indoors, to a distance of 20m. The standard has moved on through 802.11a, b, c, g, n and 802.11ac, which was introduced in Decemebr of 2013. A quick look at a local computer store’s range of WiFi modem/routers shows devices using the G standard, the N standard and the AC standard. While there are several standards beyond these, they often have poor range and haven’t been well adopted.

802.11ac devices operate in the 5GHz band, but individual devices may have backwards compatibility with 802.11n/g/b which operate in the 2.4GHz range hence these devices require multiple antennae, up to 8. The 5GHz band is often divided into 2, creatively referred to as 5GHz-1 and 5GHz-2. Devices in this range can cost upwards of $700. These devices often use very high performance processors to maximise performance.

WHY IS THERE SUCH A RANGE OF DEVICES AVAILABLE SIMULTANOUSLY?

This really boils down to cost, coverage and number of antennas. At the lower cost end we generally find 802.n devices. These are backward compatible with 802.11g and 802.11b devices. The standard specifies a coverage of 70m. Devices in this range may be dual band, meaning they can transmit and receive on two different frequency bands, being 2.4GHz and 5GHz. This is like watching two TV channels at once; they may well have up to five antennas, being say three for 2.4GHz and two for 5GHz.

COVERAGE

It’s probable that you’ve used WiFi and also probable that you’ve found poor or intermittent coverage. The farther from the modem/router you are, the weaker the signal strength. Obstacles like walls reduce the signal strength dramatically; for example, your wall may be say 15cm thick, but if you are diagonal from the modem/router it is effectively much thicker and reduces the signal even more.

There are several techniques to help overcome coverage issues. You could move the modem/router; being closer to the outside (e.g. near a window) may improve the quality of the signal it receives. WiFi repeaters are devices that receive the local WiFi signal and retransmit it at a higher level. Some modem/router models have optional high-gain antenna (HGA). Ethernet over Power (EoP) could be an alternative too; this will provide a wired network connection.

SECURITY

Let’s face it, computer security is not everything it could be and generally not everything it should be. There have been several security standards used so far:

WEP (WIRED EQUIVALENT PRIVACY): Uses a relatively short encryption key among all users and uses it often. Cracking of this key can take less than one minute, which is why WEP is not recommended.

WPA - (WIFI PROTECTED ACCESS): Introduced in 2003 as a temporary solution to the poor security of WEP, it uses a unique 128 bit key for each data packet. A security flaw has been found in WPA due to leftover problems with WEP.

WPA2: Features strong AES (Advanced Encryption Standard). A minimum of WPA2 certification is required for all WiFi devices.

WPA2-PSK (PRE-SHARED KEY): Also referred to as WPA-PSK (pre-shared key) mode, this is designed for home and small office networks and doesn’t require an authentication server. Each wireless network device encrypts the network traffic by deriving its 128-bit encryption key from a 256-bit shared key. This key may be entered either as a string of 64 hexadecimal digits, or as a passphrase of 8 to 63 printable ASCII characters. If ASCII characters are used, the 256-bit key is calculated by applying the PBKDF2 key derivation function to the passphrase, using the SSID as the salt (prevents two systems from using the same key, even though they may share the same password) and 4096 iterations of HMAC-SHA1. WPA-Personal mode is available with both WPA and WPA2.

WPA2-ENTERPRISE: designed for enterprise networks and requires a RADIUS authentication server to provide additional security. RADIUS (Remote Authentication Dial-In User Service) is a network protocol used to provide centralised management of authentication, authorisation, and accounting.

WPA3: In January 2018, WPA3 was announced as a replacement for WPA2. The new standard uses 192-bit encryption and individualised encryption for each user.

WHERE IS WIFI GOING?

There are several upcoming WiFi developments including:

• Miracast provides high-definition content sharing over WiFi. It would seem that Google’s Chromecast has this covered.
• WiFi Voice allows transmission of high quality voice signals over WiFi.
• WiFi HaLow is a low power, long range WiFi incorporating 802.11ah technologies and is intended for use by sensors and wearables.
• WiFi Direct enables WiFi devices to connect directly, making it simple and convenient to do things like print, share, sync and display. WiFi Direct devices can connect to one another without joining a traditional home, office or hotspot network.
• WiFi TimeSync allows tightly coupled time coordination between WiFi devices.
• WiFi Location provides meter-level, geo-location accuracy. This is generally the domain of GPS, and will continue to be where WiFi access is not available.
• WiGig operates in the 60GHz frequency range, which is not yet congested. This allows multi-gigabit transfers, which is ideal for multimedia streaming.

There are proposed standards offering even faster transfer speeds. At the moment, and for some time, there has been a real problem with crowding due to numbers of WiFi-enabled devices. Just how WiFi will cope in 10 years’ time, is anybody’s guess.

Ethernet

Ethernet is now a very common high-speed data transfer mechanism. But where did it come from, where is it now, and where is it going?

Ethernet was invented by Xerox at their Palo Alto Research Centre (PARC). PARC has a history of major developments including laser printers, the graphical user interface (GUI), the mouse, e-paper and unicode.

THICK ETHERNET

The initial implementation of ethernet was called “thick ethernet” because it used a cable that was 9.5mm in diameter. It was also known as “10Base5”, where “10” referred to “10Mbit transmission”, “base” stood for “baseband” (basically one frequency on or off), and “5” for “500m length of cable”. It had one inner conductor inside a thick plastic sheath, then a braided copper layer formed the signal return path, and finally a plastic outer coating. Due to the width of the cable and its copper sheath it was difficult to bend, but was popular for long distance runs. It could span 500m in a run, and handle 100 connections. In a multi-floor building it could form a backbone to connect each floor together. Thick ethernet used screw-on terminators at each end of the cable. These terminators were 50Ω resistors mounted on a connector for attachment to the cable. An interesting tool was used to connect to thick ethernet, called a vampire clamp. As it sounds, a vampire clamp had teeth that cut their way to the center conductor and attached permanently to the cable.

THIN ETHERNET

The next phase of ethernet was “10Base2” or “thin ethernet”. This used a much thinner flexible coaxial cable, again with a single inner conductor a thinner plastic insulating layer, a copper braid, and an external cover. 10Base2, much like 10Base5, can be broken down to “10” for “10Mbps”, “base” for “baseband” and “2” for, well incorrectly “200m cable run”. The correct maximum cable run (per segment) is actually 185m, but 10Base1.85 sounds a little silly. Each piece of thin ethernet cable had a BNC socket at each end, with a BNC connector. The ends of the cable were terminated with 50Ω resistors mounted in a BNC plug.

To connect to a node — being a workstation, printer, hub or switch — a T-piece was inserted between two lengths of thin ethernet cable.

While thin ethernet segments are limited to 185m, four segments could be joined together via repeaters. A repeater was generally a PC with two network interface cards running specific repeater software.

It probably isn’t obvious but the use of T-pieces to connect workstations means each workstation actually had one cable to one side of the T-piece, and another on the other side of the T-piece. To add another device to the network cable meant breaking the whole network cable, installing two lengths of cable with a T-piece, and reconnecting it all back together. Not a big deal except that on a live network you had just a few seconds before data loss occurred and 20 angry office staff would yell abuse! At the time, this messy cabling wasn’t generally considered a nuisance, but it’s far from the office norm today. The in-and-out configuration of thin ethernet was necessary as the cable was effectively one continuous signal path. It was a bus topology and quite distinct from say, Arcnet, which had a star configuration.

Another interesting fact is that ethernet still uses CDMA/CD. Another acronym? Well, what would the world be without them? Remember that we may well have 100 workstations connected to the single cable. Some control mechanism is required here because if Fred is sending data, then Max will have to wait. The protocol used is CDMA/CD, which stands for “carrier-sense multiple access with collision detection”. “Carrier-sense” means that network devices had to determine if the cable was already carrying a signal. “MA” means “multiple access”, meaning that every connected device has access. “CD” stands for “collision detection”. But hang on, didn’t we say it uses carrier sense to determine if the cable is already being used? That’s true, but it is possible that while we are preparing to send our data, another device goes ahead and sends its own data. If we then send our data it could well result in data corruption and no data reaches it’s intended target at all. If this is detected the data is “dropped”. To try to avoid further collisions, each device waits a random time before attempting to transmit again.

Many years ago I provided local technical support for the company distributing a range of equipment and software from Corvus Systems. They had a networking product called “OmniNet” that ran on twisted pair cabling. The first version had a transfer speed of 1Mbps. The second version ran at 4Mbps. In direct speed comparisons of the then 10Mbps ethernet, 4Mbps OmniNet gave around the same data transfer rates. That’s 4Mbps versus 10Mbps. Why? OmniNet didn’t use CDMA/CD. It used CDMA/CA, where “CA” stands for “collision avoidance”. Instead of pretty much throwing data onto the cable, OmniNet was really careful to ensure it had the cable to itself, therefore avoiding the associated data losses of dropped packets and the random delay before retrying.

As a trainee at the PMG’s Department (Post Master General, a government department which handled mail and telephony plus telegraphy), I was expected to believe that copper wires could only transfer electrical signals between 300Hz to 3.4kHz. This is enough bandwidth to transmit voice signals, as this is what the PMG was about, the telephone that nearly every house had. Of course now, we have 100Mbps going to many homes via the NBN. It seems that copper was quite underestimated in those days.

Modern Ethernet Standards

10BaseT: The term “10BaseT”, in fact, covers several sub-standards. While thin ethernet had a massive penetration in networking, it was a little clumsy. 10BaseT cable is quite flexible as it uses four thin insulated copper wires and a thin outer sheath. Internally, the “wires” are arranged as two pairs twisted together. You probably know that a current running through a conductor generates a magnetic field around itself and conversely, a magnetic field interacting (“cutting”) a conductor will induce an electric current. We have four wires carrying a current and each inducing magnetic fields and current into the other conductors. Twisting two conductors together means that the magnetic fields tend to cancel each other out. It should be noted that each of the two pairs is twisted at a different rate. Different twist figures further decreases “crosstalk” or the interaction of one conductor to the next. 10BaseT can have a cable run of 100m, transfers data at a rate of 10Mbps, and uses Cat3 or Cat5 cabling. The subject of cable categories is covered later in this article.

100BaseT: As the name implies, 100BaseT transfers data at a rate of 100Mbps. Like many versions of ethernet, electronics manufacturers had several attempts at making ethernet work well at 100Mbps over twisted pair cabling. 100BaseTX has become the norm. It uses two twisted pairs like 10BaseT, but 100BaseT requires a minimum of Cat5 cable.

1000BaseT: You guessed it! 1000BaseT transfers data at 1000Mbps, or to put it more simply: 1Gbps. It uses a minimum of Cat5 cable and uses four twisted pairs.

10GbE: There are several sub-classes of 10GbE. 10GbE devices generally have a socket that can accommodate different hardware modules that handle 10GbE in different ways. These modules are around the same size as a USB flash drive. Differences between standards come about due to the maximum range (1m to 80km), the type of cable required (some can operate over older cable installations), and such. 10GbE in some forms can operate over copper wires, with distances anywhere between 1m and 55m.

ETHERNET WITH OPTICAL FIBRE CABLES

There are several standards for using ethernet with fibre optic cables. The differences are mainly in the maximum distance they can reliably transmit a signal. These vary from 300m to 45km.

Let's not confuse this with Fiber Channel (note the spelling difference), which is not for general purpose data transmission. Fiber channel is a control system designed to use optical fibre cabling, instead of copper wiring between high speed storage devices and servers, and provides block level access to storage media.

Higher data transfer rates? Standards exist for 2.5, 5, 10, 20 and 45Gbps still on twisted pair cabling. 100m distance is attainable for up to 10Gbps, while the higher rates may fall to a 35m range. That range may well be very useful for connecting fast devices in a server room.

DATA PACKETS AND MAC ADDRESSES

Data transmitted over ethernet is divided into packets. A packet not only includes data, it also includes the source and destination addresses of the data, an error detection mechanism, plus other housekeeping “bits”. A packet can have from 46 bytes (also known as octets) to 1500 bytes of data. Every ethernet device has a unique 48 bit (6 bytes) MAC address. The MAC address is used for the source and destination addresses inside the data packet. The first three bytes of the MAC address represents the manufacturer of the ethernet device. You can view this info at http://standards-oui.ieee.org/oui.txt

HUBS AND SWITCHES

In the days of thin ethernet a hub was a common device. Like modern network switches, they had several ports, or connectors for network cables. They were designed to duplicate incoming traffic over each of their ports. Commonly, hubs had 5, 8, 16 or 24 connection ports. As ethernet marks each data packet with the destination address, sending the same data to many devices is clearly a wasteful concept as each receiving station has to look at the packet and decide if it’s meant for them.

A network switch is a very different device. A network switch learns the addresses of devices on its ports. In this way, if Fred wants to send a file to Max, the switch or switches know the best route to send the data, and send it via that port only. It should be noted that while ethernet is basically a point-to-point system, it can generate broadcasts which go to all connected devices. A broadcast may be initiated by BOOTP or DHCP. BOOTP can be used to find a network device that can provide a pre-configured O/S to install on a workstation. DHCP allows network devices to obtain a network address.

Network switches are available in the usual 5, 8, 16 or 24 port configurations. Network switches can generally be stacked or interconnected to function as a larger capacity switch. Enterprise level switches are generally modular, allowing multiple 10/100/1000BaseT connections, fibre connections, modular power supplies and similar. They may also have remote monitoring/management facilities.

CABLE CATEGORIES

“Category 5” or “Cat5”, was commonly used in many thin ethernet installations. It supports a transmission distance of 100m, and a minimum of Cat5e is recommended for new installations.

Cat5e: Supports a transmission distance of 100m and a bandwidth of 100MHz.

Cat6: Suitable for 10/100/1000BASE-T installations with a run length of up to 100m.

Other categories exist such as Cat7, Cat7A and Cat8. Cat7 and Cat7A while defined have not yet been adopted. Cat 8 is designed for data centre users, so we makers are unlikely to encounter it.

Power Over Ethernet (PoE)

Power over ethernet is very handy for getting DC power to remote equipment such as WiFi hotspots, IP cameras and commonly, VoIP (Voice over Internet Protocol) phones. There are a couple of standards but not all vendors stick to these. It would be wise to do checks on PoE cabling you encounter, before connecting devices.

ETHERNET OVER POWER (EoP)

Your router is in the lounge room but your main computer is in the study with the network printer. What do you do? A simple solution is Ethernet over Power. This is usually provided by a device that plugs into a standard mains power socket and has provision for a standard ethernet cable to be plugged into it. Duplicate this at the far end and your problem is solved!

THE FUTURE

It seems likely that Ethernet over Copper (EoC) or Ethernet over Fibre (EoF) will be around for a long time to come, with ever-increasing bandwidth and longer distances; however, the rise of WiFi is commonly replacing ethernet in the home and small office.