康普的 COVID-19 客户和合作伙伴中心 访问
Every 60 seconds, the world downloads more than 38,000 hours of music on Spotify, streams nearly 70,000 hours of Netflix video and sends 150 million emails. As the Internet of Things (IoT) and all that it encompasses –automation, artificial intelligence, smart buildings and more – takes shape, these numbers will only grow.
With businesses and consumers demanding more data at faster speeds, hybrid cloud architecture is becoming a necessity in the data center. Cloud computing, however, requires a fundamental shift in network design. How do enterprises, service providers and multitenant data centers design infrastructure with agility and scalability?
To deliver maximum scale and performance in a shared resource environment, data centers are migrating from traditional three-tier switching architecture to a leaf-spine architecture. Leaf-spine supports the high volume of data that is constantly being generated within the data center as many servers collaborate to serve applications. It also provides additional availability that three-tier switching does not, thanks to all-to-all (mesh) network connectivity. Now applications can be paired with any server, while redundancy provides easy server additions and migrations.
The process by which leaf-spine architecture is designed and deployed is critical to the successful implementation of cloud-based applications. While much of the attention has focused on the architecture necessary to enable cloud computing to flourish and grow, there is a very real cabling infrastructure beneath the network that connects the physical layer upon which the cloud model depends. Point-to-point cabling can quickly become unmanageable in high-speed applications. As leaf-spine networks require more high-speed optical network links, a structured cabling approach ensures scalability and can lower a data center’s costs in the long run.
East-West Network Design
Whereas the traditional, three-tier network is optimized for north-south traffic, hybrid cloud architecture is optimized for east-west traffic flows, or server-to-server traffic. Any cloud environment requires high bandwidth between servers; therefore, the network’s ability to move massive amounts of traffic between distributed servers becomes critical. East-west network design equips the data center for high capacity and greater security, as modern applications typically involve heavy traffic (as much as 10X) between servers within the data center.
To ensure application performance meets user expectations, both network availability and latency must continue to improve. So, too, must network management. Between ever-increasing fiber densities, complicated point-to-multipoint connections and heavily meshed leaf-spine architectures, managing connectivity has become increasingly challenging. Automated infrastructure management tools, such as the imVision platform, helps to track every connection and provide real-time information that data center managers can use for continuous improvements to the network.
For example, the imVision platform makes it easier to ensure capacity is available when assigning duplex and parallel optic paths. Conversely, it also assists with identifying surplus cabling and switch ports and making them available for parallel or duplex migration as new higher-speed applications become available.
While moving to a hybrid cloud architecture is undeniably challenging, integrating the public cloud with the private cloud and on-premise computing can help your data center scale up with speed and reliability. CommScope’s high-speed migration platform makes the transition to leaf-spine more manageable. Clearly, the need for speed isn’t slowing down. We’re here to help customers keep up with the demand, while getting ahead of the trends that most affect data center design and cloud computing. Learn more here.