What is a Network?
In technology terms, a network is a set of two or more devices, that are connected to each other in some fashion, and can pass information between each other. As mentioned, a network can be as simple as two connected devices. But more often than not, there are hundreds or thousands of connected devices, like you might have in a medium to large sized business. And let’s not forget the Internet mentioned above. There are literally millions of devices spread all over the world connected through various means including cables, radio waves, microwaves, infrared light, and fiber optics, to name a few!
How will you Classify a Network?
There many different ways to classify a network. Here are a couple of examples:
- Large or Small – classification is done based on the number of devices.
- Local or Wide – classification is done based on the geographic area the devices cover.
- Wired or Wireless – classification is done based on the medium used to connect the devices.
There are others, but you get the idea. Notice that in each example, classification is done based on some characteristic of the network. It follows then that a network can be classified in more than one way.
DEFINITION OF A HETEROGENOUS NETWORK
A Heterogeneous Network (as opposed to a Homogeneous Network) is a network where the devices are made by different manufacturers, or the computers run different operating systems. For example, land-line phone networks in your home that support devices from various manufacturers, or a computer network that is primarily made up of Windows-based machines, but also supports Linux machines. Its easy to see that many variations can exists. You, yourself, may have seen Apple networks with Windows machines on them, or Linux networks with Epson printers. Both are quite common in the business and academic environments.
In computer networking, a heterogeneous network is a network connecting computers and other devices where the operating systems and protocols have significant differences.
Examples of heterogeneous networks
For example, local area networks (LANs) that connect Microsoft Windows and Linux based personal computers with Apple Macintosh computers are heterogeneous. The word heterogeneous network is also used in wireless networks using different access technologies.
Heterogeneous Communication System
This case study has been proposed in the the Combest project by EADs Innovation works Germany. It concerns a distributed heterogeneous communication system (HCS) providing an all electronic communication infrastructure, typically for cabin communication in airplanes or for building automation. HCS contains various devices such as sensors (video camera, smoke detectors, temperature, pressure, etc.) and actuators (loudspeakers, light switches, temperature control, signs, etc.) connected through a wired Ethernet network to a central server. The server runs a set of services to monitor the sensors and to control the actuators. The devices are connected to the server using network access controllers (NACs).
The architecture and functionality delivered by HCS are highly heterogeneous. The system includes different hardware components, which run different protocols and software services ensuring functions with different characteristics and degree of criticality e.g, audio streaming, clock synchronization, sensor monitoring, video surveillance, etc. Moreover, HCS has to guarantee stringent requirements, such as reliable data transmission, fault tolerance, timing and synchronization constraints. For example, the latency for delivering alarm signals from sensors, or for playing audio announcements should be smaller than certain predefined thresholds. Or, the accuracy of clock synchronization between different devices, should be guaranteed under the given physical implementation of the system.
Complete details of the case study can be found on the Combest site We have developed a structural model of HCS using BIP. At top level, the structure of the model follows the natural decomposition into physical elements e.g., server, network access controllers and devices are the top-level components. Moreover, these components are connected and interact according to the wired network connections defined in the original system. Then, one level down, every (physical) component has a functional decomposition. Inner sub-components provide features for network operation (e.g., packet delivery, filtering, routing, scheduling, …), protocols (e.g., clock synchronization) or services (e.g., audio/video streaming, event handling, etc.)
The overall complexity of this case study is extremely high. A model for a relevant functional subsystem required approximately 300 atomic components and 1900 connectors in BIP. Almost all atomic components have timed behaviour. They totalize approximately 250 clocks variables to express all timing constraints. Moreover, the use of large domain data (e.g., packet numbers) and complex data structures (e.g., FIFO queues of packets) made the state space of the model extremely huge. One single state needs approximately 400 bytes to be represented. Furthermore, the state space has a heterogeneous structure which prevents its compact representation using symbolic techniques based on BDDs.
We have been interested to verify the clock synchronization protocol i.e., the protocol used to synchronize the clocks of all devices within the system. The challenge is to guarantee that the protocol maintains the difference between a master clock (running on the server) and all the slave clocks (running on devices) under some bound. A first major difficulty is network communication which makes all applications interfering and therefore requires exploration of the whole model. A second difficulty comes from the time granularity i.e., one microsecond, needed to perform faithful observations. These two factors significantly restrict brute-force simulation approaches: 1 second system lifetime needs approximately 10 minutes simulation time with microsecond precision on the BIP model.
To overcome these difficulties, we proposed in  a new verification technique which combines random simulation and statistical model checking. We have been able to derive exact bounds on clock synchronization for all devices in the system. We also computed probabilities of clock synchronization for smaller values of the bound. Being able to provide such information is of clear importance, especially when the exact bounds are too high with respect to user’s requirements. In particular, we have shown that the bounds strongly depend on the position of the device in the network. We also estimated the average and worst proportion of failures per simulation for smaller bounds i.e., how often the clock synchronization exceeds the given bound on some arbitrary run.
FUNCTIONS OF A HETEROGENOUS NETWORK
A heterogeneous network contains interconnected nodes and links of different types. Such interconnected structures contain rich information, which can be used to mutually enhance nodes and links, and propagate knowledge from one type to another.
Heterogeneous Network Advantages
NEC on Heterogeneous Networks
Networks are never uniform because network environments are never the same. Each network is likely to include densely-populated metropolitan areas requiring differing degrees or types of public and business services. Networks may also cover more sparsely-populated rural regions, or a popular tourist destination or industrial center. Networks may also experience newly-identified dead-spots where users cannot get a signal, especially at cell boundaries.
NEC’s intelligent Heterogeneous Network (HetNet) Solution is designed to address these diversified needs. NEC’s HetNet combines NEC’s high-capacity Small Cells with a multivendor-capable self-organizing network (SON) introduced into an existing network. NEC recognizes the need for flexibility, so our HetNet Solution can add cells of various sizes to existing networks with the same or separate spectrums. By doing this in the most effective and cost-efficient way, we can help you boost the available capacity in certain locations and provide uniform coverage in areas which are lacking.
Benefits of NEC’s HetNet
The main benefits of NEC’s HetNet solution are enriched capacity and guaranteed spot-free coverage. The overlay of Small Cells onto existing networks ensures high-speed, high-capacity communications for specific hot-spot area. This is especially desirable in densely-populated areas and business districts that need to provide reliable communication services to support an increasingly diverse and data-heavy range of applications. Our own studies show that HetNet can boost capacity and system throughput many times more than a conventional network.
Small Cells can also be placed over gaps between cells to fill dead-spots, or added along the network rim to expand the network. A uniform coverage is increasingly important in today’s world with so many content-based applications and services. The ultimate aim of LTE is to provide tailor-made, high-quality, spot-free network provision, but there is still some way to go. NEC can help make this aim more achievable by combining Small Cell overlay with NEC SON suites that automatically manage interference between cells and other parameters and help optimize the quality of LTE services.
NEC’s HetNet can also help maximize the cost efficiency of introducing a new network by providing cells of appropriate sizes from NEC’s plug-and-play All-in-one Small Cell eNodeB to larger cells. This flexible choice of cell size allows LTE services operators to enhance or expand their networks with minimum investment. The eNodeB’s simple site installation, automatic self-configuration and efficient network integration enables LTE service providers to implement quick yet error-free network upgrades.
NEC’s SON for HetNet
Drop your comment