Edge Network Expandability and Scalability

DOI : 10.17577/IJERTV10IS120131

Download Full-Text PDF Cite this Publication

Text Only Version

Edge Network Expandability and Scalability

Gourav Bansal

Kurukshetra University

Abstract- Edge technology moves stowage, processing, plus communication amenities from the remote repository to the edge networks, leading to reduced delay and immediate access. When it comes to performance, the confluence of edge technology and IoT with compute unloading provides a viable alternative. Calculation unloading conserves energy, decreases time complexity, and improves the battery life of resource constrained IoT gadgets, among other benefits. Whenever a significant number of IoT gadgets approach the periphery for compute unloading operations, unfortunately, edge computing confronts a scaling difficulty. To solve the scalability problem in edge technology, this article proposes two alternatives: EEFI (energy-efficient framework for the IoT) as well as EERC (energy-efficient recursive clustering) algorithm that prioritizes jobs based on weight.

Keywords: edge computing, IoT, EEFI, EERC


    Edge technology is a new computing model that has drastically transformed the technical scenery of today's collaborative apps as well as IoT-based intelligent devices. Portable gadgets like laptops, tablets, and phones to perform data, calculations, and communication are becoming more popular. These devices' processing capacity is growing gradually, but they can still contest with desktop computers. Furthermore, these devices are constrained in terms of resources, and their performance in a standard cloud context has resulted in large latencies.

    One of the significant features of edge technology is computation offloading, in which IoT gadgets transfer an intensive computing operation to an edge repository for processing [1]. However, compute unloading is a complicated process that needs knowledge of what to offload, when to unload, and how to offload. Since it experiences execution plus computing delays along with network delays, using a computation offloading technique requires careful consideration. Low latency is a strict requirement for IoT- based systems. When constructing an effective compute offloading strategy, this must be taken into account.


    A significant number of IoT devices are used to create these intelligent systems. Multiple offloading requests are often sent to the edge server by these devices. This causes congestion on the edge server, resulting in a scalability issue. An edge server's scalability refers to its capacity to handle a high count of computing unloading processes concurrently. Nonetheless, scalability has numerous drawbacks, including overcapacity backhaul networks, wasting resources, and degrading an edge server's efficiency. EEFI (energy-efficient framework for the Internet of things) and EERC are two recommended solutions (energy-efficient recursive clustering).

    IoT infrastructure is made up of a variety of gadgets of various sorts. The supply of services to these gadgets in real-time puts a strain on the resource-constrained edge servers. These significant scale devices are grouped into smaller reasonable groupings known as clusters to answer the scalability difficulty. These are calculated using the EERC technique. Whenever an event happens, the event-driven clustering algorithm EERC groups the IoT gadgets. It has a two-stage clustering system. The proximity betwixt IoT devices is calculated in step 1 using Euclidean Distance Forming "K" bunches [2]. Utilizing distance and spacing among nodes, the iterative design of the technique additionally separated the "K" group into a "j" number of nodes. A Cluster Head (CH) is chosen based on this criterion. Round-robin (RR) sequencing is used to choose CH,

    as well as the power density of the gadgets is also correctly considered [3]. For 2 rounds, each node in the cluster receives data. When contrasted to other cluster members, the gadget with the shortest turn-around duration, highest energy threshold, and most processing power is chosen as a CH [2]. When re-grouping occurs, the CH will change, then the subsequent node that meets the requirements will be picked as the new CH. CH receives data from each gadget in the group. The CH combines the data, gives a weight value to each job, decides whether or not to discharge the work, and then unloads the operation to the BS (base station). A two- stage clustering scheme is shown in Figure 1.

    Figure 1: IoT recurrent grouping

    The real benefit of EERC is expandability since it minimizes the volume of processes routed to the edge repository, protecting it from bottlenecking. Because the CH head is changed each time, the cluster's gadgets consume less power [2]. Moreover, by avoiding a solitary point of failure, this technique extends the life of IoT gadgets. At CH, data aggregation permits only highly prioritized operations to be unloaded to the edge server, maximizing the use of backhaul networks while lowering communication overhead and increasing dependability.

    The two-stage computational unloading technique (EEFI) executes a smooth unloading procedure while adhering to the task's rigorous power and delay requirements. Firstly, the method computes whether dispatching to an edge server or local processing is more useful at the IoT tier by analyzing the parameters. Secondly, at the edge layer, the method ensures that delay-tolerant plus delay- sensitive activities are processed via the edge network within

    normal workload conditions. The technique, however, unloads the lag-tolerant operation to the web for processing when the demand is high. This method guards the edge repository from overcrowding then scales it to provide maximum performance. In this case, a client-server architecture is used for mockup, with the BS acting as the server and the mobile acting as the client. Unlike centralized methods, where the BS makes the offloading choice, the EEFI approach is utilized on the client unit, which determines whether or not to unload the work [2]. The EEFI method performs well with a fixed number of users, whereas biased randomness outperforms mobile users.


      1. Additional Storage

        Numerous businesses still view the cloud to be a files stowage resolution. Stowage is still a dynamic feature of every IT design, even if current edge technology amenities do more than simply storing files. They possess sufficient stowage volume to accomplish those demands, from storing essential records plus operating applications to safely retaining vital customer files. While a shoestring business may not require far more than just a few terabytes of storage, following a run of triumphs, it may find itself struggling to handle significantly greater workloads as well as resources [4]. Organizations may utilize cloud technology to expand their digital storage plan to match their uptime demands without paying the capital investment that comes with extending physical infrastructure, rather than maintaining an ever-growing array of hard disks that expands every time a new user is taken on. Colocation server firms make it much more accessible than ever for enterprises that have already invested in virtual servers to link their infrastructure to extra cloud storage.

      2. Additional Power

        Edge technology has transmuted IT systems by permitting even the smallest businesses to access super- powered computational power exclusively obtainable to their larger rivals. Organizations are switching to the cloud to yield the solutions that provide revolutionary profitability, whether via solid programming interfaces or advanced analytic systems [4]. Due to infrastructural restrictions, corporations were previously unable to scale up their computer capabilities quickly. Not only did they have to buy new stuff, but it took a long time to get those resources in operation and sort out the kinks. Growth prospects may have passed you by then, leaving you with a considerable expense for idle gear. They can scale up the processing power of their technology as required with cloud technology, permitting them to manage the momentary spike in traffic or ramp up capacity for long- term demand growth.

      3. More Adaptability

        In today's environment, enterprises and sectors may fluctuate significantly. One year, a perfect IT technology might be outdated the succeeding, rendering it hard for companies to adjust to fluctuating consumer needs. Edge technology enables enterprises to drastically rearrange their equipment plus workloads to meet current demands without being tied by the technology as well as resources appropriate in the past. Present issues can be handled by repurposing

        private connections as multi-cloud and hybrid cloud fittings [4]. This is predominantly vital for emerging businesses, which are antagonized with novel impediments as well as permissible obligations as they scale up their undertakings. They can make such adjustments using edge technology to ensure that their IT equipment meets the demands of the time.


    Edge scalability has certain drawbacks. Expandability in edge computing necessitates increased storage space, for example. Furthermore, edge technology poses a significant security risk due to the large volume of data. The more data processed, the greater the risk of security breaches. Furthermore, it simply analyzes data, and the expense of edge computing gadgets and scalability implementation is quite expensive. This is because the entire procedure necessitates modern infrastructure as well as workforce.


    A cyclical grouping strategy is implemented at the IoT tier to tackle the expandability and scalability challenges in edge computing, limiting the sum of processes discharged to the peripheral repository plus shielding the edge from congestion. A workable alternative is a two-stage categorized computational unloading technique that meets the task's delay as well as power requirements while accomplishing compute unloading. A perfect resource scheduling approach is used to guarantee optimal usage of edge assets. This allocates the absolute necessities of edge assets to each job that is unloaded. This strategy allows the edge server to scale much farther. EEFI improves the performance of resource-limited IoT gadgets by lowering energy usage, decreasing computing time, and extending battery life. By using a client-server design for process unloading, the approach also reduces computational unloading inefficiency. EEFI meets the needs of both delay-tolerant plus delay-sensitive operations, scaling the endpoint repository to handle the highest amount of client requests possible.


      1. L. Lin, X. Liao, H. Jin, & P. Li, "Computation offloading toward edge computing." Proceedings of the IEEE, 107(8), 1584-1607, 2019 [online]. Available: https://ieeexplore.ieee.org/abstract/document/8758310/?casa_token= iZaOMLUN2R8AAAAA:Hzpn3YDeVtM42WZWrZCz6FVD7Ssv Gbwxf1w-HVxNWN3bD-


      2. M. Babar, & M. Sohail Khan, "ScalEdge: A framework for scalable edge computing in Internet of thingsbased smart systems." International Journal of Distributed Sensor Networks, 17(7), 15501477211035332, 2021 [online]. Available: https://journals.sagepub.com/doi/abs/10.1177/15501477211035332

      3. R. T. Al-Zubi, N. Abedsalam, A. Atieh, & K. A. Darabkh, "LBCH: load balancing cluster head protocol for wireless sensor networks." Informatica, 29(4), 633-650, 2018 [online]. Available: https://content.iospress.com/articles/informatica/inf1203

      4. S. Maheshwari, D. Raychaudhuri, I. Seskar, F. & Bronzino, "Scalability and performance evaluation of edge cloud systems for latency constrained applications." In 2018 IEEE/ACM Symposium on Edge Computing (SEC) (pp. 286-299). IEEE, (2018, October). Available: https://ieeexplore.ieee.org/abstract/document/8567673/?casa_token= O655g- PBkREAAAAA:AcYBwPRf0N7bNvtMeFkOlSjaTNXq29HtS3f- DeqgRR0ZLVBXlJo2eOje8gAicyy8o_5wyx4nukhR8Iw

Leave a Reply