Building High-Quality Campus Network with CloudEngine Switches
Enterprise products, solutions & services
In recent years, emerging technologies represented by Wi-Fi 6, the Internet of Things (IoT), cloud computing, and Artificial Intelligence (AI) are maturing, accelerating the transition of campus services toward all-wireless, cloud, and intelligence. As a growing number of key services are carried on Wi-Fi networks, the focus has shifted to user experience and future network evolution, instead of just bandwidth and security of campus networks.
Wi-Fi 6, a game-changing technology, provides unprecedented levels of bandwidth and latency to ensure optimal wireless experience, accelerating digital transformation of enterprises. As Wi-Fi 6 becomes the preferred choice for enterprises, what will happen to wired switching networks?
The theoretical maximum rate of a Wi-Fi 6 Access Point (AP) is higher than 10 Gbit/s — with up to 1.15 Gbit/s on 2.4 GHz and 9.6 Gbit/s on 5 GHz — while actual test results show that the maximum throughput is higher than 7 Gbit/s. Undoubtedly, Wi-Fi 6 APs have achieved a record high bandwidth over the air interface, but if no corresponding wired network is available, the performance of Wi-Fi 6 APs is wasted. To prevent the uplink ports on Wi-Fi 6 APs from becoming network bandwidth bottlenecks, device vendors have upgraded uplink ports from gigabit Ethernet to multi-gigabit Ethernet (multi-GE) that supports 2.5 Gbit/s, 5 Gbit/s, or even 10 Gbit/s.
According to the current Wi-Fi 6 AP specifications, most Wi-Fi 6 APs deliver 5 Gbit/s on uplink ports, with high-end models offering as high as 10 Gbit/s uplink speeds. Therefore, to support 5 Gbit/s and 10 Gbit/s simultaneously, the campus access switches connected to Wi-Fi 6 APs must be multi-GE capable and are recommended to provide 40 Gbit/s uplink speeds. In addition, access switches must have 802.3bt PoE capabilities to supply power to IoT devices such as wireless APs and cameras.
In fact, deploying Wi-Fi 6 APs doesn’t necessarily require a large-scale upgrade of access switches. If the current network bandwidth requirements aren’t demanding and access switches don’t result in network bottlenecks, enterprises can still use legacy access switches without the need for upgrades. Instead, they can decide on an upgrade schedule that suits their actual requirements and future service development.
Traditional campus switches use packet switching technology, where packets are switched according to the following process. First, upon receiving service flows, an inbound line card randomly selects a switching fabric to which it then sends data packets based on a hashing algorithm. Then, the switching fabric forwards the data packets to an outbound line card, which then delivers these packets to the target end user.
This conventional approach is workable in a network with a light load. However, with the growth of user concurrency and single-user bandwidth, two significant problems emerge. First, the switching fabric is randomly selected based on the hashing algorithm, which can easily cause unbalanced loads among switch fabrics and therefore affect the overall performance. Second, when multiple inbound line cards select the same switching fabric, this switching fabric becomes overloaded with concurrent traffic, resulting in the loss of data packets. As illustrated in the following figure, the switching fabric is congested when users A and B send service flows simultaneously to user C.
Featuring an innovative Next-Gen Switch Fabric (NGSF) switching architecture, Huawei’s next-generation flagship core switch, CloudEngine S12700E is unlike any other. This switch is the first of its kind to introduce cell switching and dynamic load balancing algorithms, achieving non-blocking, lossless ultra-fast traffic forwarding at the campus core. In addition, CloudEngine S12700E, resolves the congestion and packet loss problems caused by packet switching and static hashing algorithms when traditional core switches are heavily loaded.
Huawei NGSF architecture innovatively processes service flows as follows. First, upon receiving the service flows sent by users A and B, an inbound line card slices data packets into cells, selects a switching fabric dynamically based on loads, and sends the cells to the optimal switching fabric. Then, the switching fabric forwards the cells to an outbound line card, which then assembles cells into data packets and sends them to the target end user. This innovative cell switching mechanism achieves dynamic load balancing among switching fabrics, and maximizes the switching performance. In addition, multipath transmission of cells prevents network congestion and service loss caused by overload of a single switching fabric, providing better Quality of Service (QoS) assurance.
Huawei CloudEngine S12700E has been tested and verified by Tolly, a leading provider of third-party testing and validation services. The test report shows that Huawei CloudEngine S12700E-12 supports up to 288 100GE ports, providing 57.6 Tbit/s switching capacity in total and offering 4.8 Tbit/s bandwidth per card, six times that of the industry average performance.
CloudEngine S series switches are equipped with Huawei’s next-generation programmable chips. These latest chips outperform previous generations by significantly improving data exchange performance, data analysis and processing, and application identification assurance. Built on Huawei’s Versatile Routing Platform (VRP), CloudEngine S series switches provide numerous advanced features, including wired and wireless convergence (integrated WLAN AC), free mobility, horizontal virtualization, vertical virtualization (Super Virtual Fabric, or SVF), network virtualization (VXLAN), deployment automation (NETCONF/YANG), intelligent O&M (Telemetry), and advanced security protection (NetStream and Encrypted Communications Analytics, or ECA). With a fully programmable architecture, CloudEngine S series switches accelerate the provisioning of new services and features through programming without the need for hardware upgrades, fully protecting customer investments.
Huawei is the first vendor in the industry to propose and launch wired and wireless convergence features and solutions that, to date, have been widely used and proven in medium- and large-sized networks in world-renowned universities, governments, enterprises, stadiums, as well as primary and secondary schools. Huawei CloudEngine S12700E integrates the WLAN AC function based on programmable chips to manage up to 10,000 WLAN APs and allow concurrent access of 50,000 wireless users. In contrast with the traditional separate management of wired and wireless networks, CloudEngine S12700E achieves true wired and wireless convergence, including centralized service forwarding, unified device management, and converged policy control. This convergence saves a considerable amount of network construction and operating costs compared with a standalone WLAN AC or WLAN AC cards (used with modular switches). Customers only need to deploy one switch to manage devices, users, and policies in a unified manner.
Huawei CloudEngine S series campus switches are purpose-built for the Wi-Fi 6 era by providing wired network assurance for high-bandwidth application experience. In particular, the flagship campus core switch, CloudEngine S12700E, delivers six times the industry average performance, meeting the network evolution needs of large enterprises and campuses for the next ten years.
Huawei’s intent-driven campus network solution — CloudCampus — consists of Huawei CloudEngine S series switches, AirEngine series Wi-Fi 6 products, and iMaster NCE (a brand-new one-stop management and control system). This future-proof solution was named a Gartner Peer Insights Customers’ Choice for Wired and Wireless LAN Access Infrastructure in 2020, signifying its recognition within the industry. Huawei’s CloudCampus Solution will continue to build a superfast, intelligent, and open campus network centered on service experience, helping enterprises achieve their digital transformation goals.