Encyclopedia
Intelligence Edge Computing (IEC) is the key enabler of emerging 5G technologies networks and beyond. IEC is considered to be a promising backbone of future services and wireless communication systems in 5G integration. In addition, IEC enables various use cases and applications, including autonomous vehicles, augmented and virtual reality, big data analytic, and other customer-oriented services.
- 6G
- 5G technologies
- intelligence edge computing (IEC)
- internet-of-things (IoT)
- mobile cloud computing
1. Introduction
The sixth-generation (6G) is applied to new communications networks which have evolved throughout the past few years and have incorporated various technologies, such as sensitive sensors, autonomous vehicles, immersive media, and Internet of Things technologies [1]. These technologies rely on millions of communication nodes and billions of endpoints. Additionally, they face many challenges such as deficiencies in wired networks and other privacy and security problems [2]. Consequently, the role of the 6G Network is to define the right set of network technologies required to deliver these applications. To be precise, its scope has been defined to meet the communication needs of societies until 2030.
The central theme of the 6G Network is the merging of digital and real worlds in all dimensions, shown in Figure 1. In addition, we expect to see much automation in the coming years. The sheer volume of things will work at the system level, and not in private networks; thereby requiring the coordination of intelligence distributed throughout the fabric of connection. Moreover, in 6G, the information between machines and robots will be provided in partial time units to safely support various operations [3]. Furthermore, the main characteristics of 6G include:
Figure 1. Vision of 6G.
-
Providing an efficient interaction among network’s infrastructure and applications as well as supporting emerging technologies in the market that enable digital society in 2030 and after.
-
Supporting convenient and effective binding for critical connectivity and edge computing networks which could be used by new poles with tighter limits as well as more varied limits for latency and amplitude.
-
Controlling the resources effectively and rising time awareness and moving beyond the current effort of the Internet by providing high bandwidth and new case’s communication service.
However, processing, organization, and implementation of a great number of data are considered the real challenges of 6G.
Intelligence Edge Computing (IEC) is an improvement of cloud computing technology which is deployed to give easy access to the near end-users [4]. IEC is an ETSI-defined system, which is connected over a wireless network and can provide with the cloud computing resources and IT services as well as move traffic computing from center to the edge [5]. Besides, the European Telecommunications Standards Institute (ETSI) Industry Specification Group (ISG) introduced IEC as a way of increasing the network’s edge capability of storage and intensive processing [6]. In addition, IEC’s features enable the mobility and provide with the computing and mobile communication and can exchange data anytime and anywhere (e.g., File Transfer (FT), Wide Area Network (WAN) interconnection, fax, e-mail, Internet Access (IA) and the World Wide Web). Furthermore, the wireless networks used for communications include IR, Bluetooth, W-LANs, Cellular Networks, W-Packet Data, and Satellite Communication System [7].
IEC is also considered to be mobile communication infrastructure that provides smooth and efficient communication between different mobile devices. Moreover, IEC is a design that aims to reduce bandwidth and delay by moving the required resources closer to the systems that demand them. The background to this claim is based on the promising expectation of lower capital expenditures and the potential to introduce new services, which could potentially be offered separately and could also be launched at a reduced cost. Among these new services is the potential to provide ultra-reliable and extremely low latency (URLLC) connections [6][8][6,8]. For instance, less than 5 ms or less than 10 ms, for autonomous Vehicle-to-Everything (V2X), Augmented Reality (AR ) and Virtual reality (VR) applications. Subsequently, the architecture of IEC orchestration is shown in Figure 2, in which it consists of several functional blocks deployed in each IEC object. Besides, the functions of this architecture can be divided into three parts, namely border network functions, identification functions and data processing functions.
Figure 2. IEC Architecture [9].
5G is an architecture for edge computing that has been conceptualized through the enhancement of Software Defined Network (SDN) and Network Function Virtualization (NFV) [10]. Additionally, 5G technologies promote the concept of dividing networks into network resources and network functions. In 5G technology, a centralized infrastructure can be maintained while improving wireless communication at the same time [11]. This allows service providers to build a single physical network, with the potential to take into consideration high-bandwidth applications (e.g., broadcasting) and low-bandwidth (e.g., Internet of Things (IoT)) applications with time-low-latency connectivity and internal corporate networks. Increasingly development of 5G technologies enables sophisticated applications and services such as IoT, online gaming, Augmented Reality AR, Virtual Reality VR and acceleration of intelligent video [12]. The integration between IEC and 5G provides important improvements, such as enabling data processing at the network edge to reduce latency and deliver tangible business results.
IEC’s Principles
Intelligence Edge Computing (IEC) enables the use of computing servers closest to the user instead of centralized devices far from the user. Therefore, it is characterized by fast data transfer and a significant reduction in response time for 5G networks [13]. This network technology is used in advanced digital systems such as (IoT) technologies, Virtual Reality VR video games, autonomous vehicles, cloud computing, and data protection. IEC includes some main principles such as portability, connection, interaction, and character [14] which are described as follows:
Portability: A mobile computing system is connected through devices/nodes that facilitate mobility. Such devices/nodes even have some limitations in capabilities and power supplies, but they have effective processing capacity and physical portability to work in a mobile environment [15].
Connection: The quality of service (QoS) of the network connection is defined on a mobile computing system. This ensures a high level of service availability with minimal delay and prevents certain hurdles without affecting the operation of connected nodes [5].
Interaction: Mobile edge computing system nodes (IECs) are interconnected through active data transactions for communication and cooperation.
Character: The operation of the mobile device or mobile node connected to the mobile network is summarized by the reference to a specific element; therefore, the mobile computing system is able to adopt specific technology to meet individual needs and also to obtain contextual information for each node [16].
2. IEC’S Characteristics and Benefits
Distinctive characteristics of IEC are its proximity to end-users, mobility support, and densities of geographic deployment of IEC servers. IEC enhances 5G networks because it reduces the latency and enhances the use of bandwidth. The main characteristics of IEC based on ETSI white paper [17][48] includes proximity, ultra-low latency, high bandwidth, and virtualization.
On-premises: IEC can operate in stand-alone environments (for example, IEC can operate in isolation from the rest of the network) and has access to local resources.
Proximity: Typically, IEC servers are placed near device end-users, thus IEC can capture data from electrical device users for other purposes as analysis of data and/or processing of big data.
Ultra-low latency: IEC server has limited computing power, it is usually sufficient to handle emerging compute-intensive applications in real time. IEC has the potential to shorten communication latency and propagation, making IEC a promising enabler for latency-critical 5G applications. IEC also opens up opportunities to alleviate the burden on forward and back linking and to accelerate content and service response by appropriately caching popular and relevant content locally at the network edge.
Awareness of locations: based on the closeness of proximity, IEC can make use of signal information from end users to predict locations. This becomes somehow significant for location-based IEC services.
Contextual information of Network: Featuring proximity, IEC can take advantage of real-time knowledge of radio network conditions and local contextual information to improve network and service quality. For example, contextual and real-time information can be used to improve user experience across personal services [6].
Virtualization: Virtual Multiple Access Computing (vMEC) technologies are the next generation 5G networks, which is a flexible software network that supports various Internet devices and Internet of Things (IoT). vMEC is based on Network Function Virtualization (NFV) and Software Defined Network (SDN). With the development of diversified networking applications, vMEC brings intelligence to the edge of IEC, reduces latency and increases available capacity. In addition, the proposed use of vIEC for container-based virtualization technology (CVT) as a gateway with IoT devices for flow control mechanism in scheduling and analysis methods which effectively increases the quality of service (QoS) of the application [18][49].
The ultimate goal of IEC is to provide improved, low latency infrastructure with deployment speed that can be scaled horizontally or vertically based on requirements. Based on the concept of IEC, services and content can be transported to the closest end-users and obtain more quality of service while reducing connection congestion and improving gateway interconnection costs. Cloud Edge has five unique mobile computing capabilities that set it apart in the market today:
-
Network performance: Cloud Edge has the ability to transfer 10× more performance throughput than competing alternatives: more than 200 Gbps on a single Intel Xeon server. Furthermore, linear scaling, independent aircraft, data monitoring and user management allow projects to support local communities to network resources quickly and efficiently scale on the edge of the network [19][20][21][50,51,52].
-
Flexibility: Due to the actual use case and other related business, intelligence edge computing has flexibility to deploy a centralized or distributed solution. This flexibility is critical to economies of scale—the ability that Cloud Edge provides. For example, a CSP looking to provide multiple cloud services with low latency will benefit significantly by focusing control plane functionality (according to network proximity) but deploying user-level instances in a distributed manner either in the CSP or the customer edge.
-
Divergent experiences: Cloud Edge enables CSP to deliver premium services to its customers on a per-flow basis. This is achieved with a single, integrated and highly optimized platform consisting of basic mobile network and LAN functions such as vProbe, CG-NAT, deep packet inspection (DPI), optimization and load balancing.
-
Virtualization and analytics: cloud computing service providers look to complement the various service offerings in the related businesses. Intelligence edge computing has the ability to provide a real-time network and insight into customer behavior. For instance, IECs define operational efficiencies, anticipate future demand and deliver service innovation which is considered to be a core value-addition. Cloud Edge supports audit paths including security data audits and provides these capabilities, along with real-time analytics.
-
Automation: In the inevitable decoupling of mobile networks, intelligence edge computing has the ability to automate the process of integrating enterprises, end-user services, applications and dynamically expanding network infrastructure, especially with IEC applications. Applications and tools of cloud edge automation enable cloud computing services to provide the ability to quickly adjust traffic rises and falls automatically which leads to reduce the operational costs and reduce the time needed to generate revenue.
3. IEC’s Challenges
According to the special use of IEC in commercial deployment, there are some important factors that must be taken into consideration. Primarily challenges include:
-
Network openness:
major challenges are related to mobile networks edge openness, where mobile operators work to control over the entire industry chain, and business risks from each other among equipment suppliers. Multiple services and processes:
- Several types of third-party providers such as application developers, content providers, OTT operators, and network equipment vendors work with service type creation and IEC server cluster management. All participants have to face the challenge of new business models and the value chain.
Durability and Resiliency:
- When integrating smart networks into a mobile base station, the robustness of the IEC server must be ensured and that the integration between them does not affect the availability of the mobile network.
-
Privacy and Security:
Integration of intelligence edge computing and other communication systems raise many challenges about the security and privacy of users and organizations. For instance, security threats of cyber-attacks with more of consideration about privacy protection when analyzing data of different users or parties.
Furthermore, according to survey studies about IEC, here is a list of some open issues and challenges. These challenges are categorized as open issues that need further research and investigation.
The standard protocol: IEC is a modern technology that advances through the implementation phases and requires standardization arising from the collaboration of industry and researchers across an agreed platform [7].
Effective deployment: The latency can be reduced to a minimum with optimized use of bandwidth through efficient deployment of IEC. However, it may look hard to optimize spectrum use with reliance on complex system components.
Mobility User and transparency: Providing uninterrupted services to an “always on the go” customer is another challenge in the IEC environment with a transparent migration process and platform heterogeneity.
Heterogeneity and scalability: Since high-end devices use different access technologies including 4G, 5G, Wi-Fi and Wi-Max, so the heterogeneity aspect of the smooth performance of IEC operations must be met. This also entails providing scalability to different platforms with varying numbers of users [22][23][53,54].
Availability and security: Resource availability often depends on server capacity and wireless access to ensure consistent service. Besides availability, the security of data and applications from any hacker must be provided with physical measures.
Interworking between fog clouds: there are three different aspects to consider in any end-to-end system when it comes to communication challenges for gates and/or fog nodes. Communication difference, which is the communications between the gateway/haze nodes and the cloud service (public or private) and the connections between the gateway/haze node and the edge/objects/sensor networks or the connections between the gateways/fog nodes themselves, so that they can share data without the need for a cloud connection.
Data management: The required data management capabilities include (but are not limited to) [24][18]:
-
Data normalization
, which is the assimilation, alignment, and enrichment of data from various sources (objects, devices, and sensors) into a common data model with well-understood connotations. Filter and query data
- , so apps and analytics can efficiently access and use related data.
Integration with Edge Analytics
- Because the whole reason this data is captured is the ability to analyze it, create new actionable insights, make decisions, and put these decisions into action. Converting data into different representations and formats for integration with the (IoT) ecosystem.
-
Compiling abstract data and/or metadata, as preparation for local analyzes or pushing them to cloud services [24].
In addition, many new challenges must be studied in order to create an advanced ecosystem where all players in the network (i.e., IoT users, service/infrastructure providers, and mobile operators) can benefit from advanced services. These challenges are summarized as follows [24][18].
Distributed Resource Management: Resource allocation is an important challenge to the success of IEC due to limited resources, the increasing number of applications, and the massive increase in mobile traffic [25][26][55,56]. Multi-purpose resource allocation optimization is different in different situations due to the diverse nature of applications, heterogeneous IEC servers, different user requirements/characteristics, and channel connectivity characteristics. With a large number of users, the wireless channel will be in trouble and competition among users for scarce computing resources will become extremely intense [27][57]. Although the centralized approach can deliver competitive performance, it suffers from low computational complexity and huge reporting expenses. Therefore, the central approach is not suitable for distributed IEC systems [12][28][12,58]. Additionally, there may not be a dedicated backhaul for information exchange and account offloading, and even if there is, the wireless connection can be congested due to the high burden of sharing large data [29][59].
Reliability and portability: Condensation is the cornerstone of the 5G network and is expected to reap enormous benefits. However, how to manage mobility and ensure reliability is a huge challenge in these environments. First, with several smaller servers covered, user mobility can cause frequent deliveries, resulting in service downtime issues and affecting overall network performance [30][60]. After that, users (such as vehicles) may move to new locations during the account of the period. In such a case, users may not be able to receive the mathematical result because they have already exited the service coverage for their servers. Therefore, efficient computation dump forms are essential for application completion. Moreover, the dynamic change in the number of offloading users leads to random uplink interference and variable computing resources over time [31][61]. Finally, providing reliable IEC services in mobile environments is really challenging due to the time-varying dynamics of wireless communication and user mobility.
Network integration and application portability: IEC servers can be deployed in various locations within the RAN based on specific technologies, and technical and business requirements. Thus, another important challenge is the seamless integration of IEC into the existing backbone network architecture and interfaces [17][48]. The presence of the IEC and the enabled applications should not affect the basic network and peripheral hardware standards. According to [32][38], a key component of IEC integration is the ability of IEC to interact with 5G networks in directing traffic and receiving relevant control information. Moreover, the application migration entails what are called applicability requirements. This eliminates the need for application developers to design multiple versions of different IEC systems.
Coexistence of IEC and Cloud Central: Cloud Distributed centers, with abundant computing resources, can handle big data applications at near-zero time and support a large number of users. However, distributed IEC is highly desirable because the computation at the edge of the network cannot only satisfy user requirements but also reduce end-to-end delays caused by traffic congestion and transmission delay. In comparison to the HetNet architecture, it is very beneficial to implement IEC in a hierarchical manner, i.e., user layers, terminal computing and cloud computing [33][62]. In this way, the IEC vendor also injects computing resources into small eNBs so that the advantages of HetNets can be exploited to diversify wireless transmission and spread the computing requirements [19][34][50,63]. We note that a distributed IEC may not have sufficient computing resources to handle all account requests and full reliance on the cloud poses challenges in providing critical latency services. Therefore, it is self-evident to distribute critical large data/latency accounts to distributed IEC servers while moving account-intensive and delay-tolerant tasks to the DC cloud [35][64]. The coexistence of a distributed IEC and a cloud core is an important issue and more research is needed for their interactions.
Coexistence of human-to-human traffic and IEC traffic: Integrating both traditional human-to-human (H2H) traffic (for example, voice, data, and video) and IEC traffic in the 5G network is a challenging task due to the massive paired IoT connections With the various quality of service requirements and the unique characteristics of the IEC movement [11]. For example, the Internet of Things system consists of human-type devices (HTDs) and machine-type devices (MTDs) that may run different types of applications, for example, MTD with sensors and smart homes, and HTD with video games. While MTDs have a mixed set of QoS requirements, such as latency, reliability, and energy efficiency, HTDs typically require a high-speed rate with a limited energy budget [36][65]. Likewise, the IEC system must be designed in a way that meets the QoS requirements of H2H traffic while preserving the unique characteristics of M2M traffic (for example, real-time response and context awareness).
