Prodotti, soluzioni e servizi per le aziende
With the development of enterprise digital services, traditional ICT cannot meet service development requirements in terms of efficiency and cost. Enterprises need to seek new technologies to resolve problems such as increasingly complex ICT systems and low O&M efficiency during digitization. In addition, the fast rise of technologies such as the Internet of Things (IoT), Wi-Fi, cloud computing, and Artificial Intelligence (AI) is spawning new markets such as edge computing, cloud management, and Artificial Intelligence for IT Operations (AIOps), leading enterprise ICT into the data-driven intelligent era. By understanding and serving digital transformation pioneers in the industry, Huawei has summarized five trends for the future development of ICT O&M and the impact of these trends on enterprise network construction and O&M management. To keep up, we plan to release three articles — Enterprise ICT Development Trends, Enterprise Campus Network Development Trends, and Campus Network O&M Development Trends. Through these articles, we hope to help enterprise CIOs, network architects, and network management personnel with their future technical planning.
This, the first of these articles, provides an outlook on the five major development trends of enterprise ICT.
Over the past 20 years, enterprise business application systems, data storage systems, employee office systems, and enterprise management systems have been transformed from offline to online and from local disks to the cloud. This shift is not only a change to the way information is transmitted and stored, but also a change to cooperation based on the division of labor. In essence, enterprises use public cloud technology, with some services obtained through Software-as-a-Service (SaaS) subscriptions. In this way, ICT is hosted and outsourced, and services can be obtained according to the service development scale and pace. This is crucial to building a flexible and elastic ICT architecture. The subscription business model is a major feature of Anything-as-a-Service (XaaS), as well as being an increasingly common ICT investment mode. With this business model, investments become annual instead of one-off, which can further improve the utilization of enterprise funds as well as control risks. In the end, cloud and services provided by the subscription model greatly reduce the CAPEX and OPEX of enterprise ICT, realize the rapid rollout and elastic scaling of enterprise applications, and improve the flexibility and resilience of enterprise business.
Of course, cloud and XaaS are not panaceas. The enterprise cloud is developing toward a hybrid cloud architecture, constantly working to strike a balance between cost, efficiency, quality, and security. No single mode is applicable to all enterprises. Each enterprise needs to determine the proportion of different modes used in their own services based on their own service requirements.
In contrast to the concepts of Software-Defined Networking (SDN) and data centralization, edge computing emphasizes local, rapid data processing and response. This is a typical example of “spiral development” in technology due to scenario upgrades. Applications such as autonomous driving, AI-powered manufacturing, and AR/VR are sensitive to packet loss, delay, and jitter caused by network transmission, driving edge computing technologies into the enterprise ICT field. The major difference between edge computing and traditional local server processing is that although edge computing nodes are scattered across various physical areas on enterprise networks, they must be centrally managed and scheduled on the cloud.
Enterprise IoT and Wi-Fi scenarios use numerous edge computing architecture concepts and related technologies. In the IoT, huge volumes of data are collected in real time and things collaborate with each other with high precision. Therefore, the IoT is a typical scenario where edge computing is used to optimize response speed. Independent hardware is used in traditional IoT solutions as the local IoT computing gateway. With the rapid upgrade of CPUs, memory, storage space of Network Elements (NEs), and the maturity of the open environment based on Network Functions Virtualization (NFV) and containers, this hardware is expected to be integrated into NEs in the future. Competition for bandwidth between STAs as well as interference leads to unstable air interfaces in Wi-Fi environments. This, in turn, makes it difficult to detect, analyze, locate, and eliminate wireless faults on traditional networks. Traditional WLAN ACs (WACs) and Network Management Systems (NMSs) cannot solve such issues. Instead, APs must integrate AI and edge computing capabilities to perform in-depth detection and real-time analysis on the air interface environment and analyze historical big data on the cloud, ultimately helping enterprise administrators implement visualized, controllable, and easily maintained Wi-Fi.
Previously, enterprise digitalization focused on shifting the transmission and storage of information from the physical world to the digital world. Now, with the development of AI and big data technologies, the major focus of enterprise digitization will be shifted to data processing. AI and big data technologies can be used to gain insight into business opportunities of enterprise data; alleviate issues such as inefficient data analysis and risk uncertainty caused by lack of experience; and accelerate service decision-making to improve enterprise operation efficiency, therefore, Business Intelligence (BI) brought by AI and big data is a new competitive edge that enterprises are developing.
The application of AI technology depends on the following three basic enablers: algorithms, computing power, and data. We believe that most enterprises do not need to invest a lot of energy in researching AI algorithms and computing power. This is not a field that enterprises are best suited to or need to pay attention to. Instead, ICT suppliers will offer a series of products and platform services to provide powerful and numerous algorithms as well as simple AI programming suites. In this way, enterprises can redirect their attention to data, which is the key for setting them apart from others. In stark contrast to the openness and standardization of algorithms and computing power, enterprise data is a private asset that is strictly guarded. Therefore, to come out on top in the AI and big data technology revolution, enterprises must fully grasp in-depth data mining and its innovative uses. Regarding enterprise network data mining, we have two suggestions: First, telemetry should be used to gradually build a digital twin of the network and streamline the data sharing of numerous management systems, which will help enterprises greatly reduce OPE. Second, the future network will be an important source of data and an important bond between online application data and offline physical data. Therefore, enterprises must focus on products and solutions related to network big data, including Location Based Service (LBS) and customer profiling.
The core of enterprise digitalization is data, which is also the key to future business competition. Therefore, the increasingly severe security issues of the digital world will become an important focus for enterprises. In recent years, well-known ransomware attacks have awakened enterprises to the fact that the losses caused by security issues can be direct and serious. As such, enterprises are realizing that merely hardening security borders is insufficient for preventing the penetration and spreading of threats.
Border security design is based on the traditional assumption that threats only come from external networks. Security borders are marked and the traffic crossing these borders is the only concern. However, due to the wide use of various tunneling and encryption technologies, as well as the constant emergence of new Advanced Persistent Threats (APTs), border security is becoming increasingly ineffective. Once the border has been crossed, the threat will spread throughout the entire internal network.
Currently, the security industry is rethinking how to better collect, share, analyze, and apply threat intelligence, as well as how to introduce big data and AI into threat prevention using threat detection mechanisms beyond traditional signature and rule-based algorithms. Instead of building a trusted zone, the focus is shifting to building proactive threat detection and defense based on zero-trust networks.
With a bigger focus on ubiquitous security, enterprise applications are embracing the cloud, XaaS, AI, and big data innovations. Such trends are only made possible with the support of the management and operation of enterprise ICT organizations.
First of all, these trends will pose new challenges to the ICT innovation culture of enterprises. Enterprises need to be more bold in trying out new and perhaps even unproven technologies. Of course, this kind of trial should be carried out in a controlled environment; for example, POC tests for overall solutions provided by suppliers, trial use in lab environments, or large-scale operations after small-scale pilots. Such trials may affect the existing ICT organization structure, division of labor, and workflows. Enterprises need to properly manage the resulting changes and gradually transform their ICT management methods based on an “agile development” approach.
Second, these trends require a shift in focus from the features and performance of the hardware industry to the overall solution. This not only places high requirements on ICT product and service design and development capabilities, but also requires enterprises to eliminate barriers between their ICT teams and encourage in-depth collaboration between network and application teams as well as security and network teams. When comparing products, enterprises should not evaluate supplier capabilities based merely on functions and performance indicators. Instead, they should assess the ability to solve end-to-end problems according to service scenarios. To help enterprises improve production efficiency and reduce OPEX, suppliers must understand an enterprise’s services, abstract scenario-based characteristics, and convert such abstractions into easy-to-use Human-Machine Interfaces (HMIs). We recommend enterprises perform POC testing to verify and experience products and solutions.
Last, such trends also place more requirements on ICT personnel skills, and these requirements are always changing. We need to manage this digital world in a more digital way. Many enterprises still use traditional methods such as the manual, CLI-based management of ICT services. Today, the trends of service-driven networks and non-CLI-based operations are gaining steam, and a growing number of HMIs will be replaced by Machine-to-Machine (M2M) interfaces. The use of AI for making ICT management decisions will become the mainstream in the next 3 to 5 years. This requires ICT personnel to shift their focus from using and maintaining ICT products to developing an in-depth understanding of enterprise services. ICT personnel may need to learn about different vendor products and accumulate O&M experience. Such skills will be gradually transferred to AI tools. Professional ICT personnel will need to be capable of automated and intelligent programming using vendor APIs to convert enterprise business requirements directly into fully automated workflows. In the future, rather than holding expertise in just one area, ICT engineers will need multi-faceted capabilities.
