Monday, July 3, 2017

Connecting to the Internet

Dial-up

Telephone lines can be used to transmit computer data. This was how people connected to the internet in the early days. However, telephone wires are designed to transmit analog information, and computers can only consume digital information. A dial-up modem converts analog signals to digital and vice versa.

To connect to the internet with a dial-up modem, you enter a phone number for your modem to call, which is provided to you by your ISP. You also have to provide a username and password. The connection process is noisy and takes several seconds to complete. ISPs would often charge by the minute, so you never wanted to leave your connection open when you weren't using it (you also couldn't make phone calls while connected). Dial-up connections use a protocol called Point-to-Point Protocol (PPP), which is specifically designed for transmitting data over dial-up.

The unit of measurement that is used to measure the speed of data across a telephone line is a baud. The maximum speed a telephone line can achieve is 2,400 baud. As modems improved over time, they could pack more and more bits into each baud. For example, a 33.6 Kbps modem can pack 14 bits into each baud (2400 x 14 = 33,600). The highest speed that can be achieved through dial-up is 56 Kbps. Dial-up connections also have fairly high latency compared to other internet connection approaches.

To try to break the 56K barrier, some ISPs experimented with server-side compression. This involved compressing certain kinds of data before sending it over the wire to the client, resulting in higher download speeds. This approach was hugely successful for certain kinds of data that can be easily compressed, such as HTML pages and plain text. But many data formats are already compressed, such as ZIP files and streaming video, so no speed improvements could be gained from them. Image file formats like JPEG and PNG already use compression, but ISPs would compress them even more, resulting faster speeds, but, as a consequence, a loss of image quality.

ISDN

As dial-up modems began approaching the 56K limit, telephone companies began converting all their analog telephone lines to digital. The process of sending digital signals across digital telephone lines is called ISDN, and it allows speeds of up to 64 Kbps (wow!).

An ISDN line contains two types of channels. Bearer (B) channels are used for voice and digital signals and run at 64 Kbps. Delta (D) channels are used for setup and configuration data and run at 16 Kbps. A common setup would be to install two B channels and one D channel, giving you speeds of up to 128 Kbps. This setup was referred to as basic rate interface (BRI). A more powerful, but less common, setup involved twenty-three B channels (providing 1.544 Mbps) and one 64 Kbps D channel. This was called primary rate interface (PRI) or a T1 line. The main downside to ISDN connections was that you had to be within 18,000 feet of the central ISP building for it to work.

DSL

Digital subscriber line connections use your telephone line like dial-up, but the connection is always-on and is much faster. They also allow you to make phone calls while the connection is active. Speeds can vary anywhere from 3 Mbps to hundreds of Mbps. The most common type of DSL connection is asynchronous DSL. ADSL has upload speeds that are slower than download speeds. On the other hand, synchronous DSL (SDSL) gives you identical upload and download speeds, but is more expensive. Just like with ISDN, you must be within a certain distance of the main ISP office. The distance can vary from a few hundred feet to 18,000 feet.

Cable

A cable connection piggy-backs off of your cable television connection. It provides upload speeds between up to 20 Mbps and download speeds of over 100 Mbps.

Fiber

There are two kinds of fiber connections. In fiber-to-the-node (FTTN), the ISP installs a central box somewhere in your neighborhood, which is connected to the actual fiber line. Then, the individual houses connect to the box using standard Ethernet or coaxial cabling. In fiber-to-the-premises (FTTP) your house is directly connected with the central office via fiber. Fiber varies in speed, but can be as fast as 1 Gbps (which is what Google Fiber provides). In some cases the download speed matches the upload speed. I have an FTTP fiber connection that gives me 100 Mbps upload and download speeds.

Satellite

The main benefit to a satellite connection is that it works anywhere in the world. No infrastructure is required (telephone lines, cable lines, etc). A satellite dish must be professionally setup so that it has line-of-sight communication with the satellite up in space. The main downsides are: higher than average latency and signal degradation in cloudy weather.

References

Sunday, July 2, 2017

Wireless Communication Protocols

There are many different kinds of technologies that allow for the wireless transmission of digital information through the air. These include Wi-Fi, Bluetooth, infrared, and cellular.

Wi-Fi

The most well known is the 802.11 family of protocols, more commonly known as Wi-Fi. In a typical Wi-Fi setup, all computers connect to a central device called a WAP (wireless access point). WAP is the technical term for “wireless router”. Every wireless network has a service set identifier (SSID), which is a human-readable name for the network. The SSID is what appears when you search for available wireless networks in your device's Wi-Fi settings. The WAP broadcasts the SSID so new devices can find and connect to the network.

For small buildings, like a SOHO (small office/home office) environment, only one WAP is needed because its signal can reach all or most parts of the building. This is referred to as a Basic Service Set (BSS). However, larger buildings cannot make due with just a single WAP. In this situation, multiple WAPs are strategically placed throughout the building, and are joined together into an Extended Basic Service Set (EBSS). In a EBSS, all the WAPs have the same SSID, so as you roam around the building, your device automatically switches WAPs based on whichever has the strongest signal.

Securing your WAP

Hiding the SSID: It's possible to configure a WAP to not broadcast its SSID, which helps prevent unauthorized people from accessing it.

Enabling MAC address filtering: Every computer device has something called a MAC address, which is a 48-bit, globally unique identifier. You can provide your WAP with the MAC addresses of all your devices so that no other devices are allowed to connect.

Changing the admin password: Many WAPs leave the factory with identical administrator passwords. Change it! The administrator password is used to access the configuration settings of the WAP (usually through a web interface), so it's important to have a strong and unique password.

Controlling physical access: Many WAPs have a handful of Ethernet ports on them. Connecting a computer to one of these ports bypasses all the wireless security that is in place, so you should either disable these ports or place your WAP in a location that only authorized personnel can access. Also, when you buy internet service for your home, the ISP often provides you with a WAP that has the Wi-Fi and administrator passwords stamped onto the case. So if you don't want to change them, make sure the WAP isn't in a place that can be seen by strangers (like your window sill!).

Ad Hoc Mode

Connecting to a wireless network through a WAP is referred to as “infrastructure mode”. But it's interesting to note that a WAP isn't required to network computers wirelessly. In “ad hoc mode” (also sometimes referred to as “peer-to-peer mode”), computers connect directly with each other to form an Independent Basic Service Set (IBSS). This is useful if a WAP isn't available and the number of computers you need to network is small.

Antennas

The antenna most commonly used by WAPs and computer devices is a dipole antenna, which is a type of omni-directional antenna. They look like a stick but actually have two antennas inside them. Some WAPs have detachable antennas, which gives you the option of installing larger, more powerful ones.

Signal strength (called “gain”) is measured in decibels (dB). Most WAPs broadcast at around 2 dB, and some let you adjust this. You might think that the higher the gain, the better, but not always. Lowering the gain to an amount that just barely covers your building will prevent your neighbors from being able to connect to your network. This also does your neighbors a favor because it lowers the amount of RFI (radio frequency interference) that their wireless networks will have to contend with.

The orientation of the antenna matters. This is called polarization. If an antenna is standing straight up, it has a vertical alignment. If it is laying flat, it has a horizontal alignment. Since the antenna in your laptop is located in the lid next to the screen, it generally has a vertical alignment when the lid is open. In order to communicate effectively, the antennas of the computer and the WAP must have similar polarities. It's good practice to tilt the WAP's antenna to a 45 degree angle to accommodate the largest variety of polarities.

Wi-Fi Security Protocols

Because all communication is traveling through the air, anyone with the right equipment and skills can intercept this communication and read it—just like tuning your car radio to a radio station. Unlike radio broadcasts, the information that travels through Wi-Fi networks can be very sensitive. To help protect your privacy, various security protocols have been released over the years.

WEP. Created in 1997, this protocol encrypts all communication with 40- or 104-bit encryption. And it was not very secure. For one, it uses the same encryption key to encrypt all communication with all client computers, which makes it possible for a single computer to listen in on everyone else's communication. And in 2001, a serious encryption flaw was discovered which allowed a WEP key to be cracked in minutes. WEP was officially retired in 2003 and replaced by WPA.

WPA. This protocol corrects WEP's weakness of using a single encryption key by changing the encryption key for every packet of data that is transmitted (called TKIP). The encryption key size was also increased to 64- or 128-bits. And it includes a feature which prevents malicious clients from altering and resending data packets. WPA was only intended for temporary use until the WPA2 standard was finalized.

WPA2. Finalized in 2006, WPA2 includes all of the improvements that WPA brought to the table, as well as an improved encryption algorithm called AES. AES is a very strong algorithm that no one has been able to find a significant flaw in (yet). In fact, the U.S. government approved it to be used for transmitting classified information in 2003. WPA2 is currently the most secure wireless security standard, and it's what all your devices should be using. WAPs that support “mixed-mode” allow devices to connect using either WPA or WPA2 (for older devices that do not support WPA2).

WPS. What if you want to connect a device like a printer or scanner to your Wi-Fi network? Because these devices often lack display screens, how are you supposed to give it the SSID and password of your Wi-Fi network? Enter WPS. It allows you to connect a device to a network with as little as two button presses. First, you press the WPS button on the device. Then, you press the WPS button on the WPA (your WPA must support WPS). And bingo, it's connected. However, it has a major security flaw. It also allows you to connect devices to it using an eight-digit code, which an attacker could use to brute force his way into the network. Therefore, security experts recommend that you turn WPS off if your WAP supports it.

Sidenote: HTTPS

You might be nervous about transmitting sensitive information over a wireless network, especially if it is a public Wi-Fi network, like the one at Starbucks or your favorite coffee shop—AND YOU SHOULD BE! Even if the network uses the best possible encryption standard (WPA2), not only could someone theoretically discover a flaw at any time and start intercepting your data, but the owners of the WAP could theoretically configure their WAP to intercept and log all information that travels through it! Or, attackers could set up their own WAP within range of the legitimate WAP and configure their WAP to broadcast an SSID which is identical to that of the legitimate WAP, causing your device to connect to the attacker's WAP instead of to the legitimate one (if I recall correctly, this was done at the 2016 Olympics in Rio).

However, you need not worry as long as you are browsing secure websites (using the HTTPS protocol) and using apps that use secure connections. The encryption standard that protects you is called SSL. When using this standard, your computer encrypts the data before sending it over the air. What’s more, the data can't be decrypted until it reaches its intended recipient. So even if someone intercepted your communication, they wouldn't be able to make any sense of it because it is encrypted. God forbid if someone breaks SSL—the internet as we know it would grind to a halt, because this standard is what makes possible such things as online shopping and online banking!

The 802.11 family of protocols

A number of different Wi-Fi protocols have been released over the years, each of which have different characteristics. These are the low-level protocols that the security protocols discussed above run “on top of”. I'll refer you to my Computer Networks 101 blog post for a description of these protocols.

Bluetooth

For short-range, wireless communication, Bluetooth is often used. It is designed to do very specific things and is not intended to be general purpose, like Wi-Fi is. A Bluetooth network is called a PAN (personal area network). It is extremely resistant to RFI (radio frequency interference) due to the fact that it hops frequencies about 1,600 times per second.

Every Bluetooth device is assigned a “class”, based on its range. Lower class devices use less power because they don't have to transmit as strong of a signal.

Class 1 100 meters
Class 2 10 meters
Class 3 1 meter

Many different versions have been released over the years (summarized in the table below):

Version Max speed Description
1.1, 1.2 1 Mbps

2.0, 2.1 3 Mbps A feature called Enhanced Data Rate (EDR) improves its max speed.
3.0 + HS 24 Mbps The high speed (HS) feature is optional and uses a Wi-Fi network to achieve the full 24 Mbps bandwidth.
4.0, 4.1, 4.2
“Bluetooth Smart”
24 Mbps Focuses on power consumption, security, and IP connectivity.
5.0 24 Mbps Focused on the “Internet of Things”, aims to be low power.

Infrared

Infrared is most commonly used in remote controls, like the one for your television. But it can also be used to transmit digital information. The Infrared Data Association (IrDA) protocol uses infrared light as its communication medium. However, it is very limited. It only supports speeds of up to 4 Mbps and is half-duplex. And it only has a max range of 1 meter. Plus, it relies on line of sight communication (any physical object placed in its way will break the link). Because of these limitations, IrDA no security features—why bother make any when the computers have to be so close to each other and it’s so easy to block the signal? Note that some computers have what looks like an infrared receiver, but these are usually used for remote controls, not for IrDA.

Cellular

Lastly, we have cellular. Cellular data connections are often referred to as 1G, 2G, 3G, or 4G. These do not refer to specific standards, but are loose terms that refer to how recent and fast the underlying technology is. At the present time, the fastest cellular technology is LTE. It is considered 4G and theoretically supports speeds of up to 300 Mbps download and 75 Mbps upload.

If you are not in range of a Wi-Fi network, you can tether your device to your cell phone. My understanding of this is that you can download apps that do this, but you need to jailbreak your device in order for them to work.

References

Saturday, July 1, 2017

The Power is Yours!

Every computer has a box called a power supply, which is responsible for supplying electricity to the internal components of the computer. Its main task is to convert the AC (alternating current) power from the electrical outlet to DC (direct current) power, and then dole out the DC power to the computer's internal components. Different parts of the world use different voltage standards for their electrical outlets, so a power supply has to be compatible with the voltage standards in your part of the world. For example, power outlets in North America run at around 115V, and those in Europe generally run at around 230V. Some power supplies have a physical switch on the outside that tell it what voltage to expect (called fixed-input). Others will adjust automatically (called auto-switching).

Due to the nature of AC power, power supplies can take damage over time from something called harmonics. Harmonics is caused by the way in which electrical devices draw power from an AC connection, and is what causes electrical devices to make faint humming sounds. Most power supplies come with circuitry that protect against this, called active power factor correction (active PFC). You should never buy a power supplies that does not have this.

I need more power, Captain!

Every power supply has a maximum amount of wattage it can draw. If the internal components of the computer try to draw more than that, the computer won't work right. For example, if you want to install a brand new, high performance graphics card, you should make sure your power supply has enough available voltage. Note that power supplies are replaceable, so if your current power supply isn't good enough, you can always replace it.

Power supplies do not use all of the AC power it consumes. Some power is lost due to inefficiencies and released in the form of heat. Most power supplies are at least 80% efficient, and they will advertise what their efficiency is on the packaging. A more efficient power supply will consume less power.

It's important to note that power supplies only draw the amount of energy that is actually being used by the computer—they do NOT draw the maximum amount they are capable of. For example, if you have a power supply can that provide a max of 500 W and your computer is only using 200 W, then the power supply will only draw enough power for 200 W. You won't be wasting electricity if you buy a power supply that can supply more power than your computer needs. In fact, it is good to have a such a power supply for two reasons: (1) To allow room for future upgrades and (2) to account for the fact that power supplies produce less wattage over time due to wear and tear.

Rails

The DC power that the power supply generates is doled out through three voltage rails. Each rail supplies a different voltage: 12V, 5V, and 3.3V. The 12V rail is typically used to power devices that have motors of some sort, such as hard disk drives and optical drives, but there is no restriction regarding what each voltage rail can be used for (for example, a high-end graphics card might want to use the 12V rail).

Each rail has a maximum amount of amperage it supports, and this is monitored by circuitry called over-current protection (OCP). Single-rail systems have a single OCP that monitors all the rails. Multi-rail systems have one OCP per rail to monitor each rail. If the amperage in any rail is exceeded, the power supply will shut itself off to prevent damage to itself.  When multi-rail systems were first introduced, they were very unstable due to poorly written specifications, but they have gotten much better since then.  For computers that use a lot of power, like servers and gaming PCs, multi-rail systems give your system extra protection against short-circuits.  For an ordinary, low-wattage desktop PCs, it doesn't really make a difference whether you have a single-rail or multi-rail system.

Power supply standards

Various power supply standards have been released over the years. ATX (also called ATX12V) introduced the idea of providing a constant supply of power (5V) to the motherboard, even when the computer is off. This is called soft power, and it allows the computer to implement various power saving features. This is the reason why you always should always unplug a computer before servicing it! This standard was later improved upon by subsequent standards (below).

ATX12V 1.3 added the P4 connector, which supplies extra power to the motherboard. It also added the AUX connector. The downside to this standard was that it was not specific enough, which resulted in power supply manufacturers producing wildly different power supplies.

EPS12V was created for servers that need more power than the average desktop machine. It added a 24-pin motherboard power connector. It also introduced the idea of “voltage rails” (explained above).

ATX12V 2.0 adopted many of the advancements that EPS12V brought to the table. Notably, it added a 24-pin P1 connector and voltage rails.

Connectors

Many of the different connectors you will see coming out of a power supply are listed in the table below. Yeah! Tables!

Connector Voltages Pins Description
P1 power connector 3.3V, 5V, 12V 20/24 The older variant of this connector has 20 pins. The newer variant (which is backward compatible) has 24 pins and provides more current.

Molex 5V, 12V 4 Typically used to power storage devices, like hard drives.

Mini 5V, 12V 4 This connector used to be used for 3.5” floppy disk drives and isn't used much anymore. You have to be careful when plugging in this connector because it is easy to plug in upside down, which will ruin the device.

SATA power connector 3.3V, 5V, 12V 15 Only used for SATA hard drives. In practice, only the 5 V and 12V voltages are used.

SATA slimline connector 5V 6 A smaller version of the SATA power connector.

SATA micro connector 3.3, 5V 9 Even smaller!  Can't reliably find a photo of this one.
P4 connector 12V 4 Used in conjunction with a 20-pin P1 connector to supply the motherboard with extra power.

AUX connector 3.3V, 5V 6 Also used for supply the motherboard with extra power.

EPS12V
EATX12V
ATX12V 2x4
12V 8 This connector goes by many different names. One half is compatible with the P4 connector.

PCIe Connector 12V 6/8 In some 8-pin connectors, two of the pins are detachable so make them compatible with the 6-pin version. It looks similar to the EPS12V connector, but is not compatible with it.


References

Sunday, June 25, 2017

Computer Networks 101

In your typical white-collar work environment, each employee has a computer at their desk. The computers are connected with each other over a LAN (local area network). If you have internet access at home, the computers in your house are most likely organized into a LAN as well. A LAN is a group of computers that are physically close to one another and that can communicate with each other over a network.  All of the computers in a LAN are said to belong to a broadcast domain, which means that if one computer sends out a broadcast message, then all the other computers can hear it.

Ethernet

One of the most common ways to join computers into a LAN is to use Ethernet. There are four properties of an Ethernet cable: EMI resistance, heat resistance, flexibility, and speed.

EMI resistance: STP (shielded twisted pair) cables are designed to protected again EMI (electromagnetic interference). A shop floor is a good example of a place where STP cables should be used because it has lots of electrical motors and other machinery. However, the vast majority of environments do not require significant protection from EMI, so they use UTP (unshielded twisted pair) cables, which are less expensive.

Heat resistance: Ethernet cabling is often run through the walls and ceilings of a building (called plenum space) in order to keep the cables out of the way. These areas of the building can get very hot. The rubbery outside of an Ethernet cable is typically made with a material called PVC. If PVC starts to melt due to high heat, it can give off poisonous fumes. A plenum-grade cable, however, will not melt in the heat because it is made out of material that is designed for heat-intensive environments. Plenum-grade cabling is much more expensive than PVC, so you should only buy it for cabling that you intend to use in plenum space.

Flexibility: You also need to think about the kind of physical wear-and-tear the cable will be getting. Will the cable sit in plenum space, untouched for most of its existence? Or will the cable spend most of its time in your office drawer, being used for various purposes around the office? Standard core cabling is made out of material that is flexible, which means you can bend it, step on it, and twist it (to a reasonable extent) without breaking it. Solid core cabling, on the other hand, is not so flexible. But its advantage is that it is a better conductor and will transfer data more effectively.

Speed rating: Every Ethernet cable has a speed rating, which defines its max data transfer speed. A cable’s speed rating is usually stamped on the outside of the cable itself. It’s sometimes referred to as a “CAT rating”, since the speed rating begins with the letters “CAT”. The maximum cable length varies between speed ratings, but for most speed ratings it is 100 meters. The ratings are measured in Mbps (megabits per second) or Gbps (gigabytes per second). To get a better feel for how fast this is, I like to divide this number by 8, which tells me how many bytes per second it supports. Note that, in order to take advantage of the full speed a cable offers, all other parts of your network infrastructure must support that speed rating, such as the network cards in the computers and the switches.

Standard Max speed/notes
CAT 1 This is the technical name for a telephone cable! Telephone cables use a RJ-11 connector, whereas Ethernet cables use a RJ-45 connector.
CAT 3 10 Mbps, some variants support 100 Mbps
CAT 5 100 Mbps
CAT 5e 1000 Mbps
CAT 6 1000 Mbps, 10 Gbps (55 meter max cable length)
CAT 6a/e 10 Gbps
CAT 7 10 Gbps with better shielding

Switches

All the computers in an Ethernet network connect to a central device called a switch, which routes the various network data to where it needs to go. This is called a star bus topology—“star” refers to the fact that the computers connect to a central switch (instead of to each other) and “bus” refers to the central device that routes all traffic.

Note that a device called a “bus” can serve as the central device as well, but buses are much more inefficient than switches because they broadcast all messages they receive to all computers, whereas switches only send out messages to the computer that the messages are intended for. Switches used to be more expensive than buses, but not anymore.

To prevent unauthorized computers from connecting to the network, you can disable unused ports on a switch.

Structured Cabling

Larger companies have the money and talent to organize their networks using structured cabling system. The aim of such a system is to create an organized, secure (both from an information safety perspective and a physical safety perspective), and reliable way of connecting all of your company’s computers to each other.

A typical structured cabling system is organized as follows. All cabling, including Ethernet cables and telephone cables, are run from each work area (the office space that an employee occupies) to a central room called the telecommunications room. This cabling is referred to as the horizontal cabling. Each piece of horizontal cabling is referred to as a run. Vocabulary rocks!

In an ideal environment, the horizontal Ethernet cabling would run through plenum space and be of plenum-grade, solid core construction. Each work area would then contain wall outlets that connect to the horizontal cabling. It’s interesting to note that Ethernet wall outlets have CAT ratings as well! Therefore, it’s important to make sure the outlet matches the CAT rating of your horizontal cabling.

The telecommunications room is the central destination for all the horizontal cabling. It contains specially designed equipment racks which are used to store its computer equipment. All rack-mounted equipment adheres to a measurement standard, simply referred to as U, which defines the height of the equipment. 1U equals 1.75 inches. Most rack-mounted equipment is either 1U, 2U, or 4U.

One piece of equipment you’re likely to find in a telecommunications room is a patch panel. A patch panel makes it easy to rearrange your network without having to mess with the horizontal cabling (which often uses fragile, solid core cables). The horizontal cabling is plugged into the back of the patch panel using a connector called a 110 punchdown block. This kind of connector connects the individual wires inside of the Ethernet cable to the patch panel. A punchdown tool is used to attach the cable in this way. Connecting the horizontal cabling to the patch panel is a time consuming process and is meant to be more or less permanent. The other side of the patch panel contains much more flexible RJ-45 ports, which are easy to plug and unplug (kind of like the telephone switches of old). Patch cables are plugged into these ports. Patch cables are short (typically 2-5 feet long), standard core, UTP Ethernet cables. You then use the patch cables to rearrange your network as you like, as often as you like.

SOHO, Bro!

SOHO environments (small office/home office) do not always have the luxury of implementing a structured cabling solution. But there are a number of technologies that you can use to form a LAN without this.

What’s the Wi-Fi password?

The most common and quickest way to create a LAN is to go wireless. Wireless networks are not as fast as wired networks, but for most purposes, they are fast enough. Various wireless standards have been released over the years, each of which varies in speed. In general, they are backwards compatible with each other. Most wireless routers support multiple standards anyway, so you don’t have to worry too much about compatibility most of the time.

Standard Max speed Frequency Range
802.11b 12 Mbps 2.4Ghz 300 feet
802.11a (came out after b) 54 Mbps 5 GHz 150 feet
802.11g 54 Mbps 5 GHz 300 feet
802.11n 100+ Mbps 2.4 & 5 GHz 300+ feet
802.11ac 1 Gbps 5 GHz 300+ feet

Since Wi-Fi transmits its data over the air, securing your Wi-Fi network is of the utmost importance. The latest wireless security protocol is WPA2—all the other standards are vulnerable to security flaws, so you should never use them. Your wireless network should also be password protected, otherwise anyone can connect to it. Another way to secure your network is to configure your router to disable its SSID broadcast, which is what causes your network to appear on a device’s list of available networks. You can also enable MAC address filtering, which only gives pre-approved devices access to the network. Lastly, you should change the router’s administrator password because routers are often all configured with the same administrator password when they leave the factory.

One downside to Wi-Fi is that the wireless signal can be disrupted in many ways. Thick or metallic walls in your building can weaken or stop a wireless signal. Any devices that use the same parts of the wireless spectrum can cause interference as well, such as baby monitors and garage door openers (this is called radio frequency interference or RFI). If you have neighbors that have their own wireless networks, they can interfere with your network too. The parts of your building that get weak or no signal are called dead zones.

Ethernet over Power


If Wi-Fi isn’t an option for your particular environment, you can buy special devices that plug into your electrical outlets which allow you to create an Ethernet network using the electrical wiring of your house! This is called Ethernet over Power (not to be confused with Power over Ethernet, which supplies electrical power through an Ethernet network). This is an example of a bridge because it connects two dissimilar network technologies. Ethernet over Power only supports speeds at around 100Mbps however, so it’s not very fast.

Sunday, June 11, 2017

A Primer on IP Addresses

Just like your have a home address that uniquely identifies your residence out of all the residences in the world, computers have IP addresses, which serve the same purpose.  They uniquely identify a computer in a network so that it can receive messages from other computers.

IPv4

IPv4 was created when the internet was born in 1981 and is still used today.  It is the network communication protocol that computers use to talk to each other over the internet.  An IPv4 address is a unique identifier that is used to identify an individual computer that is connected to the internet.  It is 32-bits long and is commonly represented in dotted-decimal notation.  This notation divides the bits into four, 8-bit chunks and displays each chunk as a number ranging from 0 to 255.  Each number is separated with a dot.  For example: 192.168.2.1.

At its inception, the set of all possible IPv4 addresses, called the address space, was divided into “classes”.  Each class contained a finite number of “chunks” of addresses.  The number of addresses in each chunk varied depending on the class.  The idea was that institutions, such as companies and schools, could purchase one of these chunks, and then dole out the addresses in the chunk to all the computers on their network.  Larger institutions with lots of computers could purchase a more expensive, higher class chunk that had lots of addresses, while smaller institutions that had fewer computers could purchase a cheaper, lower class chunk that had fewer sub addresses.

The classes are summarized below.  If you want to learn more about the logic behind how they were organized, I suggest you read this Wikipedia page.


Class
Number of chunks
Number of addresses in each chunk
Class A
128
16,777,216
Class B
16,384
65,536
Class C
2,097,152
256
Class D
reserved
Class E
reserved

Do you see a problem here?

The problem with this scheme was that companies were unlikely to use every address that was available to them.  The choices for the number of addresses you could have varied wildly—you could have 16,777,216, 65,536, or 256!  You couldn't have anything in between!  If a company needed, say, 1,000 addresses, they had no choice but to purchase a Class B address and put all the rest to waste.  To top it off, some of the organizations that were involved in the early development of the internet possessed Class A chunks, which they were hardly making any use of.

This started to become a pressing issue as the internet grew.  The risk that all IP addresses would be used up, called IP address exhaustion, became a real possibility.

CIDR

As shown, the way the class system divided up its chunks of addresses was very coarse-grained, which resulted in lots of wasted addresses.  To combat this, the class system was done away with in 1993 and replaced with a system called CIDR (Classless Inter-Domain Routing).  This system gives organizations many more choices regarding how many addresses they are assigned, which results in less wasted addresses.

CIDR uses something called variable-length subnet masking (VLSM), which allows the address's subnet mask (the part that identifies which organization an address belongs to) to be of any size.  The class system, on the other hand, only permitted the subnet mask to be 8 bits (Class A), 16 bits (Class B), or 24 bits (Class C) long.  With CIDR, if your company only needed 1,000 addresses, you could purchase a 1,024 chunk (22-bit subnet mask, leaving 10-bits for the address, 2^10=1,024).

CIDR notation consists of an IP address, followed by the number of bits the address uses for its subnet mask.  For example, 192.168.100.14/22 represents the IP address 192.168.100.14 with the first 22 bits of that address being the subnet mask.

But CIDR is only a stop-gap measure.  The IPv4 address space consists of about 4.3 billion addresses, which seems like a lot.  But on a global scale, it is not.  If the internet continues to grow, the IPv4 address space will soon run out.  A more permanent solution would be to increase the length of the IP address.  Enter IPv6.

IPv6

Created in 1998, IPv6 addresses are a whopping 128 bits long, resulting in an incredibly large address space of 3.4 x 10^38 (the number of grains of sand on Earth...or something?).

IPv6 addresses are represented as eight, four character, hexadecimal strings separated by colons.

FEDC:0000:0000:0000:00CF:0000:BA98:1234

Because they are so long, there are tricks you can employ to make them shorter.  If a segment contains all zeroes, you can replace the segment with a single zero:

FEDC:0:0:0:00CF:0:BA98:1234

If an address contains consecutive segments which consist of all zeroes, you can replace them with a double colon (but you can only use this trick once):

FEDC::00CF:0:BA98:1234

And if a segment begins with zeroes, you can leave the zeroes out (unless the segment contains all zeroes, in which case you must leave one zero in):

FEDC::CF:0:BA98:1234

IPv6 and IPv4 are not compatible with each other, which complicates the migration process.  While it is likely that the network card in your computer supports both IPv4 and IPv6, the infrastructure around the globe that makes the internet work cannot switch over so easily.  It will be a long and piecemeal process.  But if all goes well, you won't even know it happened.