Organizations around the world doubled down on digital transformation at the start of the pandemic, and that innovation continues to deliver advantages. Data is unlocking insights and driving operational actions, enabled by accelerated adoption of AI and intelligent edge computing. Augmented/virtual reality, 5G wireless, and SD-WAN open the door to breakthrough applications and digital experiences.
We’re only just beginning to imagine the outcomes from creating, analyzing, moving and storing unimaginable amounts of data. Global traffic volumes show no signs of slowing.
Continued innovation across the compute, storage, and network domains is essential to drive the next wave of digital transformation not only in centralized locations but also at the edge, where the work gets done. Infrastructure must keep pace as the industry evolves from data centers to “centers of data.”
400GbE is here, and it’s going to be huge
It wasn’t so long ago that 10GbE seemed like an unfathomable amount of bandwidth to connect servers and storage. Then 25GbE and 40GbE took its place, and soon enterprises were aggregating 100GbE links to deliver the performance necessary for data-driven applications.
The adoption of more efficient protocols such as NVMe puts additional pressure on the network. A server with four NVMe drives can quickly saturate a 100Gbps link.
100GbE has been the most widely deployed speed in data center switching since 2019, led by the cloud. Looking specifically at the enterprise portion of data center switching, 100GbE had the most robust growth in 2021 as enterprises shifted to higher speed ports to support AI/ML and edge-based workloads where traditional 10/25Gbps networks cannot keep pace (Source: 650 Group).
The momentum in server access isn’t slowing. 650 Group forecasts that 25GbE and 100GbE ports will comprise 73% of the controller and adapter market by 2024. Server designs will not only rapidly shift to 100GbE, but also will also see multiple 100GbE ports per server in high-end workloads. The analyst firm projects bandwidth growth to accelerate in the middle of the decade.
It’s not just about speed. Data center network architectures have evolved as well, moving to a leaf-spine topology for greater scale, responsiveness, and agility both in data centers and at the edge. An EVPN-VXLAN network fabric extends layer 2 connectivity over the physical network, creating a more agile, secure, and scalable network.
Consumption models are changing too, with many organizations embracing a hardware-as-a-service model to ensure that innovation is not constrained by a refresh cycle driven by capital depreciation.
HPE Aruba Networking 400GbE data center switch addresses these trends
We are thrilled to introduce the HPE Aruba Networking CX 9300-32D data center switch to meet these rapidly evolving trends.
The HPE Aruba Networking CX 9300-32D is a next-generation 12.8Tbps, 1U fixed configuration switch with 32 ports of 400GbE that supports port breakouts for 200G, 100G, and 25G[1]. The switch is designed to provide flexible, cost-effective, high-density 100/200/400GbE for servers, storage, and intra-fabric connectivity.
[1] 25G breakout is a future item.
The CX 9300-32D switch protects enterprises’ investment as they transition server farms from 10GbE to 25GbE and 100GbE/400GbE EVPN-VXLAN spine configurations, with reduced power and a smaller footprint. It supports large data center pods of up to 6,000 x 25G servers or up to 2,000 x 100G servers. This is an 8X jump in scaling/density over the current HPE Aruba Networking CX 8325-32C, which scales to 700 x 25G servers.
The CX 9300-32D can be used as a 25/100G leaf or 100/400G spine switch (128 x 100G ports). When deployed as a spine switch, it can connect to a range of leaf switches, including CX 8325-48Y8C, CX 8360-48Y6C, CX 10000-48Y6C Distributed Services Switch, or the 9300-32D to connect 100G, 200G or even 400G servers.
Operators can also leverage HPE Aruba Networking Fabric Composer to manage these increasing large and complex networking fabrics. HPE Aruba Networking Fabric Composer greatly simplifies the automation, orchestration, and end-end visibility of the fabric underlay and virtualized overlay, helping to accelerate provisioning across rack-scale compute and storage.
HPE Aruba Networking CX switching data center use cases
This 400GbE announcement comes on the heels of HPE Aruba Networking’s continued integration into dozens of HPE solutions spanning hyperconverged , mission critical, HPC and emerging Ethernet Storage Fabric use cases. All these solutions are highly complementary of HPE Aruba Networking’s recently announced Distributed Services Switch, the HPE Aruba Networking CX 10000 with Pensando, a new category of switch that delivers stateful software-defined services at 100X the scale, 10X the performance at one-third the TCO of traditional network and security solutions.
Design for today, plan for the future with HPE Aruba Networking
HPE has long been an industry leader in compute, virtualization, and storage, while HPE Aruba Networking is a proven leader in networking. Our combined experience across the full stack gives us a unique vantage point to help enterprises make adopt breakthrough infrastructure innovations to realize their digital strategies. An agile, scalable, and high-performance foundation can change the game, allowing enterprises to deliver unparalleled digital experiences to customers and performance with unprecedented operational efficiency.
Learn more
- Data center modernization
- HPE Aruba Networking Ethernet-based storage fabrics bring cloud-like scalability to data-driven applications
- Delivering faster business outcomes
- CS 9300 Switch Series at-a-glance
About the Author
Head of Data Center and Core Switching PLM, Global Product Line Management, HPE Aruba Networking, Vishwas Shetiya is passionate about product management, marketing and engineering, with more than 15 years experience in data center and enterprise networking, encompassing diverse products, technologies, and business units.
