China
Africa
Explore chapters and articles related to this topic
Edge AI
Published in Pethuru Raj, Anupama C. Raman, Harihara Subramanian, Cognitive Internet of Things, 2022
Pethuru Raj, Anupama C. Raman, Harihara Subramanian
Digital data generated by digitized entities and edge devices are being aggregated and subjected to various investigations to emit usable intelligence to arrive at intelligent devices. Having AI capability at edge devices, the widely anticipated real-world and real-time data analytics can see a glorious reality. ML models can be created in traditional and large-scale cloud environments and used in edge devices to infer new data items. Sufficiently powered by in-memory databases and AI, edge devices can set the ball rolling for visualizing and realizing next-generation sophisticated applications. With 5G communication capabilities and the proliferation of edge clouds, last-mile connectivity is becoming simpler and speedier. Newer possibilities will emerge with edge computing. Image classification, face recognition, object detection, machine vision and intelligence, outlier identification, etc., can be facilitated at the edge.
Wireless Networks
Published in Jiguo Yu, Xiuzhen Cheng, Honglu Jiang, Dongxiao Yu, Hierarchical Topology Control for Wireless Networks, 2018
Jiguo Yu, Xiuzhen Cheng, Honglu Jiang, Dongxiao Yu
The promotion of WMANs is to meet the growing market demand for broadband wireless access (BWA). Although 802.11x technology has been used and obtained great success in BWA for many years with many other proprietary technologies, the overall design of the WLAN and the features are not well suited for outdoor applications of BWA. When it is used outdoors, the bandwidth and the number of users are limited, and there are other issues such as communication distance. Based on this, the IEEE decided to develop a new and more complex global standard that should solve the problems of both the physical layer and quality of service (QoS) aspects to meet the BWA and the needs of the “last mile” access market (Figure 1.4). The last mile refers to the portion of the telecommunications network chain that physically reaches the end-user’s premises.
Technology Primer
Published in Dan Rayburn, Streaming and Digital Media, 2012
While streaming media does increase security and enable delivery of large files, it also imposes a significant delivery cost to the network. As shown in Figure 2.1, the Internet infrastructure of the content delivery value chain can be segmented into four separate regions. The First Mile includes all the equipment where the information is stored, such as databases, servers, data center infrastructure, and access connections. The Internet Cloud, or Middle Mile, refers to the vast interconnect networks that make up the Internet, such as thousands of service provider transport pipes, routers, and switches, and other invisible peering points of the Internet. The Internet Edge is actually not a real “edge,” but refers to the point at which a home consumer or business user connects to the Internet at large. The Last Mile denotes the local access connection between the user and the Internet Edge.
Challenges and Viability of Use of PLC for Personal Communication in Underground Coal Mines
Published in IETE Technical Review, 2021
R. N. Raul, S. Palit, T. Maity
The narrowband PLC (NBPLC) system works using the CENELEC bands which is in the frequency band from 145.3 kHz to 500 kHz. The transmission data rate of NBPLC is in the range of 50 kbps which is quite low. The NBPLC systems are applied in smart systems like smart metering, lighting, energy and grid management, etc. The frequency bands for the broadband PLC (BBPLC) are from 1 MHz to 250 MHz. The transmission data rate is 10 mbps which is quite high [20]. Due to this data rate the application of BBPLC systems are like Last Mile Telecom (LMT), internet, high definition TV (HDTV), etc.
A Survey on Packet Switching Networks
Published in IETE Journal of Research, 2022
The structure of the end-to-end WAN connections comprises two stages; first, the WAN access technologies are employed in the local loop (last mile). Second is the packet-switched network (IPS/carrier WAN technologies) in the cloud, see figure 1.
A Complete Detection and Mitigation Framework to Protect a Network from DDoS Attacks
Published in IETE Journal of Research, 2022
Ram Charan Baishya, D. K. Bhattacharyya
In a DDoS attack the attacker first compromises a set of machines, called as bots(zombies). The entire network of such compromised machines controlled by a single user (i.e. the actual attacker) is called a botnet. Once control over one or more botnet is achieved, the attacker commands the zombies to send Internet traffic to one or a set of selected servers called as victim(s) in an attempt to exhaust the resources of the victim such as CPU, memory and link bandwidth (the link could be any where between the first mile and last mile router). Under such a situation the legitimate users of the victim(s) experience high degradation of service or at worst situation, no service at all. Figure 1(a) shows a pictorial representation of a DDoS attack along with its participants such as attacker, bots and victim. However, it is not always necessary that the attacker needs to have a botnet of compromised machines to perform a successful DDoS attack. A variant of DDoS attack that doesn’t require a botnet to launch the attack is known as reflection DDoS attack (DRDoS) [1]. In a DRDoS attack vulnerable public servers which response to queries, mostly over UDP such as DNS and NTP are used as reflectors to carry out a DDoS attack. Figure 1(b) shows a pictorial representation of DRDoS attack. To perform an attack, the attacker sends requests to such reflectors but spoofs the SIP of the request packets as the victim’s IP. When the reflectors receive such requests they send the responses to the corresponding SIP, i.e. to the victim. The victim thus receives a large volume of response messages from many such reflectors sufficient to create a DoS situation at the victim. Two important advantages of DRDoS attacks are as follows A large botnet is not required to carry out a massive attack. Even with a single machine an attacker can launch a massive DRDoS attack.In DRDoS attack, the attacker can achieve a bandwidth amplification of the final attack traffic towards the victim. The size of the response messages of the reflector servers such as NTP and DNS, are typically many times bigger than that of the request messages. The bandwidth amplification can be calculated as the ratio of the number of bytes in the response messages to the number of bytes in the corresponding request message. The attacker can easily amplify the attack traffic by many hundred times using proper reflectors. The different types of reflectors and along with their amplification factors are well documented in [1].