[SOUND] Hello, welcome back to the course on web connectivity and security in cyber-physical systems. This lesson starts a module on link layer connectivity protocols and application layer communication protocols. In this video lecture, we'll talk about connectivity solutions available for embedded systems. In other words, we will focus on machine to machine or M2M connectivity protocols and their basic characteristics. Machine to machine connectivity protocols facilitate the establishment of a connection between two or more devices. Here we will be discussing the protocols that reside on the link layer of the TCP, IP model. They establish a path between two end points. For our cyber-physical system case, the connectivity solutions can broadly be divided into two main categories, wired and wireless. Wired connection are the one's that require a physical connection between endpoints. While in wireless connections, we don't have any physical medium. As we will see this classification affects the abilities of the protocols in many ways. For example range, date rate, and power consumption. Wired connections generally tend to have higher date rates and higher power consumption but this is not always the case. There are certain parameters that help us defining and distinguishing connectivity protocols. Mainly range, data rate, power consumption, topology, TCP/IP support, spectrum, estimated node support, approximate cost. Well, each of these will be discussed in this lecture. The range defines how far away or to what distance the protocol can offer connectivity. Based on this definition of range, we divide the protocols broadly into four logical categories, PAN, LAN, MAN, and WAN. PAN or personal area network interconnects devices within the personal space of an individual. The range is typically from ten centimeters to ten meters. LAN or local area network interconnects devices within a limited area such as a home, office, or building. It ranges typically from ten meters to some hundreds of meters. MAN or a metropolitan area network interconnects devices within a city or district covering several buildings, a large university network and a citywide municipal network are good examples. The range is typically from 500 meters to some tens of kilometers. WAN, or wide area network, generally has no geographical restrictions on its location or size. The WAN enables different LANs to connect to each other. Although the distance is not a limit as such, a WAN can be restricted to a specific country for example. The data rate defines how much data per second you can transfer. Or how much data the protocol can theoretically transfer per second. In wireless networks, the theoretical maximum is generally quite different from actual results. For cyber-physical systems, we divide the data rates into low, high and very high categories. Low data rates transfer up to one megabits per second. High data rates more than one to two megabits per second. And very high data rates more than 50 megabits per second. This classification doesn't apply when thinking in terms of typical human usage of networks. Remember that two megabits per second of sustained throughput is pretty high for most embedded systems, because your smart energy meter or smart lighting system, like Philips Hue, doesn't watch Netflix. Data transfers are typically periodic updates which do not require high data rates. Power consumptions gives you an idea about battery life and a data throughput of your interconnection protocol. Generally, the lower the power consumption, the lower the data rate. Higher rates mean the device is on for longer periods of time and has to process more and faster, thereby increasing power consumption. Here we divide power consumption into two categories, idle and active. Idle power consumption is the power used when the devices is in the state of rest. Meaning, no communication is taking place. Idle power is a significant factor, for example, in sensor networks. Where sensors are mostly idle communication wise. Active power consumption is the power used when a device is in the full communication mode. This will give you the battery life of your node when the node is under full communication load. Another consideration is whether the connectivity protocol and modules offer different power profiles. Such as idle state enabling profile, low data rate profile, receive only profile, and transmit only profile. Each of these save battery power in its own way. The topology refers to the interconnection between different nodes in the network. There are different ways a protocol might allow devices to connect to each other. The most common or dominant ones are the following. In a bus topology there's a central line or bus and each device is connected to that line. So logically each device is directly connected to the other devices. In a star topology we have a central node or server and each device connects to this server. In a mesh topology, each device is connected via point to point links to two or more other devices. In a full mesh topology, each device is connected to all other devices in the network. A tree topology is a mixture of the star and bus topologies. In a tree network, two or more star networks are connected by a bus network. An important question is, does the connectivity protocol support TCP/IP on the higher layers? This is essential in two main ways. First, supporting TCP/IP means easy aggressing and connectivity to existing internet or legacy systems that are already deployed. Second, it makes web connectivity very easy. In case there is no support for Internet connectivity, a gateway can be used. For example, a large set of smart tags can connect to a central server, which is connected to the Internet. In this case, the central server is the gateway to the Internet for those tags. Gateways have their benefits and drawbacks. For example, they add extra cost and complexity. On the other hand, they can be used to offer extended functions. A gateway could, for example, encrypt data coming from the tags before passing this data to the Internet. Or it can manage authentication and limited access to each tag. By doing so the gateway essentially extends and complements the services provided by the connectivity protocols themselves. It is important to know which part of the microwave spectrum a wireless connectivity protocol uses. There might be regional restrictions on the spectrum which is used by your device. Or a device might be operating in a commercial band requiring you to get a license to operate. There are different bands but for simplicity we can divide them into two categories, free and commercial. The free band is the portion of the wireless spectrum that doesn't require devices to have operating licenses. The most common one is the industry of scientific and medical or ISM band. The frequency ranges are vast but as an example Bluetooth and WiFi operate in the two point four gigahertz ISM band. An example of the commercial band is 900 and 1,800 megahertz range used by GSM. Estimated node support is simple to understand. It means the number of clients that can be connected to each other, either directly or via gateway. There might no theoretical limit on how many devices can be connected. But there comes a point where the performance get severely degraded when a lot of nodes are connect to a single gateway, for example in the case of a WiFi. The last parameter is the cost. When deploying a small set of devices a difference of some euros per device might not matter too much. But when deploying hundreds of thousands of nodes even a few cents can affect the overall price greatly. In cyber-physical systems and censor networks, the number of nodes is often large enough to make the price a decisive factor in selecting the connectivity protocol. In this lecture, we have defined the basic properties of connectivity protocols. In the next video, we will have an overview of a set of popular connectivity protocols. [SOUND]
