Ever since the first 802.11 standard was published in 1997, it has evolved to become the de facto protocol for much of the wireless networking across a wide range of devices and implementations. Today the protocol family includes 802.b 802.11a, 802.11g, 802.11n, and 802.11ac, respectively. Some of these will play a very important role in the IoE. There are other 802.xx protocols (such as 802.15) that will have a strong presence in the Internet of Everything, as well. This series of article will take a look at the technical side of these specifications.
This first discussion will address the common MAC layer inherent to all flavors. Subsequent articles will discuss the PHY layer, the various evolutions of the original specification, and other technicalities that have made it what it is today.
Generically, the IEEE 802.11 standard defines the specification of two of the layers; the physical layer (PHY), and the media access control (MAC). Figure 1 is a graphic of the full stack. This article discusses the MAC layer. Both the MAC and PHY layers are responsible for realizing wireless local area network (WLAN) communication in the standard 2.4, 3.6, 5, and 60 GHz frequency bands. The MAC layer also handles security. Originally this was used in computer networking but as the IoE evolves, this will port to many of the devices living on the IoE.
Brief background
The IEEE LAN/MAN Standards Committee (IEEE 802) is the governing body that develops and maintains these standards.
Figure 1. The 802.11 protocol stack.
802.11 is a generic family of standards for wireless networking in the Wi-Fi domain. The first version was the IEEE 802.11, which was released in 1997. It is the core spec and defines two raw data rates of 1 and 2 Mb/s). Originally the transmission medium was IR in the Industrial Scientific Medical (ISM) frequency band, at 2.4 GHz. However, it was never deployed.
The spec also defined Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) as the access method in the MAC layer. It was purposely done that way to be identical to the Ethernet protocol.
One of the reasons this spec has so many revisions is because the original offered so many choices that interoperability was a challenge. There was a lot going on at the time so the working group set a very loose standard, more of a “meta-spec,” really, with a great deal of flexibility in the parameters. Unfortunately, many vendors saw that as an opportunity to design a lock-in spec, which created a fair amount of confusion. Eventually, as technology advanced, the spec matured and evolved to what it is today.
One thing worth mentioning is that when the 802.11standard was adopted. It defined a couple of methods for basic authentication. However, one of them, open authentication, was a simply a protocol to allow Wi-Fi association with all 802.11-compliant devices. It really didn’t provide any authentication as we have come to define it to day.
A second method was Wired Equivalent Privacy (WEP). This was a bit more secure in that shared key WEP authentication actually permitted a shared key to be used for authenticating users for network access. However, by 2001, WEP was cracked. This was a major breakdown in the technology and it created a huge security breach that impeded the rate of Wi-Fi adoption in the enterprise. (It was joked that WEP stood for Weak Encryption Protocol.)
After that the group went back to the drawing board, but the breach was so bad the industry turned to the Wi-Fi Alliance, which in late 2002 released a new security standard called Wi-Fi Protected Access (WPA). It was based on the current IEEE work on 802.11i, standard and was quickly implemented to correct the problems with WEP. Eventually the 802.11i addendum was finally ratified and WPA was official.
Now it’s time to dig into the layers that deal with security, authentication and a few other jobs.
802.11 MAC layer technical discussion
There are two types of medium access defined in 802.11, a distributed coordination function (DCF) and point coordination function (PCF). DCF is used with asynchronous data transmission. Basically that is where all stations have simultaneous access to the medium using CSMA/CA (discussed a bit later), with a random back-off implementation. This is much like a bunch of cars at an intersection where each senses traffic within the intersection in a cooperative fashion.
With distributed coordination function, the wireless stations poll the state of the medium before transmitting a data packet. This is a rather simple scheme where, if the idle time of the medium is greater than a distributed inter-frame space (DIFS), then the station can transmit its packet. If not, it goes back to idle mode and continues to monitor the medium. To keep from falling into a set timing loop, once the medium goes idle, the station waits for a DIFS and then decreases the back-off counter in a step-wise manner. This sets a timer that differs each cycle and once the back-off timer expires, the station transmits its packet. With these random timer settings across the stations, there is little chance that any station will back off for any appreciable amount of time. If the transmission is successful, the receiver sends and after a short inter-frame space (SIFS).
The point coordination function implements what is referred to as, a time-bounded type of service. Such services rely on a point manager rather than letting the stations slug it out as with the distributed coordination function. In most implementations today, that is the access point (AP). In this scenario, the access point determines which station is given the go-ahead to transmit. This is the typical traffic cop scenario wither there is a controller that gives each vehicle, its turn. This particular methodology is intended for transmission of real-time traffic, which will be important as IoE traffic skyrockets. It also can be used for asynchronous data. This access method is based on a polling scheme controlled by an access point, and will find a great deal of applicability in the IoE because the number of access points will increase radically.
The design of the polling scheme is either a priority-based scheme, or just a simple round robin method. The wireless channel uses something called a superframe structure (defined as a contention-free repetition interval), which consists of a contention-free period (CFP) and contention period (CP).
At the beginning of the contention-free period, the access point will transmit a beacon frame to all of the stations in the basic service area (BSA). Next, the access point will confirm the medium is idle for a specific point-inter frame space (PIFS). The Beacon frame contains a variety of information including the beacon interval, the basic service set (BSS) identifier, and the maximum duration of the contention-free. This is a control directive from the access point, and instructs all of the basic service set stations not to send any packet in the contention-free period after receiving a beacon.
Once this has been set up, and during the contention-free period, the access point polls each station in its polling list (with either a DATA+CF, or CF-poll frame) and allows it to transmit if conditions are correct. Then, each station will respond, using one of numerous acknowledgement strings. The access point repeats the polling of each station until it reaches the maximum duration of the contention-free period. At that point, the access point generally terminates the contention-free period with the transmission of a CF-End frame. Then the process simply repeats, and communication occurs during each of these cycles.
This is somewhat oversimplified, but a visit to the IEEE will reveal a bounty of deeper data. Next we will dissect the functions of the MAC layer.
More MAC
The MAC layer also handles a myriad of other functions, from authentication to power saving, but the most interesting ones are related to security.
The authentication protocols of 802.11 come in two forms. One is the shared key authentication, the other is the open systems authentication
Open Systems works something like this: The devices’ wireless network interface card sends an authentication request frame to the access point. The access point responds by returning an authentication response frame. That frame contains either approval or disapproval of the authentication, which resides in the status code field in of the frame body. If the authentication is valid, there is a handshake, and the connection is established. If not, then there is no connection established.
This is a fairly simple and uncomplicated process. It establishes that the host and the target are who they say they are. This is the one process that is required in all communications.
Shared key authentication is a bit more complex. It is a four-step process that resolves authentication based on whether or not the authenticating device has the correct wired equivalent privacy (WEP) key. The network interface card sends an authentication request to the access point in the form of a frame. The access point responds by inserting challenge text into the frame body of a response frame and sending it back to the network interface card.
The network interface card then uses its WEP key and encrypts the challenge text, then returns it back to the access point, via another authentication frame. After decryption, the access point compares the challenge text to the initial text. If the texts are equivalent, the access point accepts that the network interface card has the right key. The access point then concludes the process by sending another authentication frame to the network interface card notifying it that it is either approved or disapproved.
Association
Once authentication has happened, the next step is association. This is required to because the network interface card must be connected to the access point that it has been communicating with. Synchronization ensures that critical parameters such as data rates are supported, and that devices cannot be captured by any other device.
The process is similar to other protocols discussed earlier. An association frame request is issued by the network interface card. It contains elements such as the Service Set Identifier (SSID) and data rates. And again, the access point responds by sending back a response association frame with information such as the association ID and access point parameters. Once all this happens the link is ready to exchange data.
WEP
The above is a generic discussion on the workings of the MAC layer. Over the years a lot of layered features were added, but to include them here would be much too lengthy and would not change the basic premise of the layer. Since WEP is part of the MAC layer security protocol, it is prudent to touch on it here, briefly.
For confidentiality, WEP uses the stream cipher RC4. For integrity it implements a CRC-32 checksum algorithm. With WEP implemented, the NIC encrypts the body, rather than the header, using a common key, of each frame prior to transmission. At the receiving end, the frame will be decrypted using the common key. The premise is simple.
A brief note on this is that with the original 802.11, the 64-bit WEP standard specifies a 40-bit key, but there is no method for key distribution. That makes 802.11-based wireless LANs relatively vulnerable to eavesdroppers. To combat this 802.11i was developed, which, in turn, developed WEP2, WEP2+ and WPA.
WPA and WPA2 are orders of magnitude more secure than WEP. WPA was simply, an interim solution, software-implementable, to allow older hardware, that did not support WPA2, to remain on line for a time.
CSMA/CA
The last component that is part of this discussion of the MAC layer is CSMA/CA. It is a network access technology that allows nodes to get on the airwaves, cooperatively. It specifies how the node uses the medium; such as when to talk, when to listen. It is the de facto carrier transmission protocol for 802.11 networks. What makes CSMA/CA so nice is that it is a preemptive protocol that prevents collisions, rather than detecting them after the fact (such as CSMA/CD).
CSMA/CA requires a delay in network activity after each completed transmission. That delay is proportional to the priority level of each device. High-priority nodes are programmed for short delays, while low-priority nodes are programmed for longer delays.
CSMA/CA is really a three-part protocol. The first technology is carrier-sense. This says that each network interface card on the network will, first “listen,” and senses whether there is any other traffic on the network.
The second protocol is multiple access. This says that all computers have access to the network simultaneously.
The third component is collision avoidance. If the network interface card senses that the network is not in use, it will send out a signal that tells the network it wants to send data. If all is still good, then the data is sent.
The combination of CSMA with CA provides improved access control over CD, simply by eliminating the all the associated overhead of retransmitting collided packets.
The process is relatively straightforward. The transmitting station first checks the medium to determine its availability. There is a short request to send the transmitted packet (RTS), which contains such information as the source and destination network addresses, and the period of the subject transmission. If the network is available, the target station responds with a clear to send packet (CTS). At that time, all other devices on the network recognize this, acknowledged claim to the network, and allow the transmission to happen, unimpeded. If something happens and the target destination does not receive an ACK packet, the source continues to retransmit RTS packets until access is granted.
There is, of course, much more involved in all of this from a technical perspective – enough to make this article interminably endless. But stay tuned. Upcoming articles will discuss the 802.11 PHY, the technologies behind the evolution of 802.11 up to its latest rendition, and why it is critical to know how this will play in the IoE.
Missive
802.11 will play a very big role in the IoE. In fact, it will likely be the most prolific wireless protocol of the future. There are all kinds of evolving Wi-Fi platforms coming, including VoWiFi, WiGiG, and simply evolutions of Wi-Fi, such as Hotspot2, which will reshape the Wi-Fi landscape. This will be especially true over the next five years as the IoE develops.
802.11 has tremendous potential in a plethora of areas within the IoE. It will be an integral component of most devices and networks going forward. With versions such as 802.11ac, and advancements in bandwidth, security and interoperability, 802.11 will be the most visible component of the global network, and integrated into virtually every chip produced.