Table of contents

• 基于“增大化现实”技术uba Reference Architecture for Data Center
  • Reference Architecture Components Selection
    • 基于“增大化现实”技术uba CX 8300 Data Center Switches
      • Spine Switches
      • Leaf Switches
      • Out-of-Band Management Switches
    • 基于“增大化现实”技术uba Fabric Composer
    • NetEdit
  • Reference Architecture Physical Layer Planning
    • Cables and Transceivers
    • 港口速度组
    • Split Ports
    • Media Access Control Security
  • Reference Architecture Capacity Planning
    • 带宽计算
    • Network and Compute Scaling

基于“增大化现实”技术uba Reference Architecture for Data Center

Aruba ESP数据中心参考架构旨在支持高可用性计算机架，使用连接在第3层脊柱和叶片拓扑中连接的冗余顶部开关。脊柱和叶片拓扑优化了性能，并提供了一种水平可扩展的设计，可以扩展以容纳不断增长的数据中心，而不会中断现有的网络组件。最初可以使用少量构建数据中心，只需两个脊柱开关即可。当最多需要额外的容量时，可以在单个面料中部署四个脊柱开关。下图显示了具有三个脊柱开关的参考架构。

Certain application environments do not require high availability at the individual compute host. In this case, a single ToR switch per rack will provide a more cost effective data center network. In this type of implementation, the number of compute hosts deployed per rack should be kept low as a ToR switch under maintenance will impact connectivity to all compute hosts in the rack. The following topology shows a single ToR design with two spine switches.

Reference Architecture Components Selection

以下部分提供了guidance for hardware selection based on your compute host, availability and bandwidth requirements.

基于“增大化现实”技术uba CX 8300 Data Center Switches

The Aruba CX 8300 portfolio offers two models of fixed configuration data center switches. The CX 8325 model offers the highest port density while the new CX 8360 model offers a variety of port configurations for small and medium spine and leaf topologies. Both models offer the following data center switching capabilities:

• High-speed, fully distributed architecture with line-rate forwarding
• High availability and in-service ToR upgrades with VSX
• Cloud-native and fully programmable modern operating system built on a microservices architecture
• Error free network configuration with software defined orchestration tools
• Distributed analytics and guided troubleshooting provide full visibility and rapid issue resolution
• Hot swappable and redundant load-sharing fans and power supplies
• 用于不同数据中心设计的电源到端口和电源到电源冷却选项
• Jumbo frame support for 9,198 byte frames
• Advanced L2 and L3 features to support VXLAN spine and leaf with MP-BGP / EVPN control plane
• Distributed active gateways for supporting host mobility

Spine Switches

The Aruba ESP Data Center reference architecture is built around two choices of 1RU high density spine switch with QSFP ports capable of 40GbE/100GbE speeds. The 8325 model can support up to 32 compute racks in a single ToR switch topology or up to 16 compute racks in a dual ToR switch topology. The 8360 model can support up to 12 compute racks in a single ToR switch topology or up to 6 compute racks in a dual ToR switch topology.

The primary function of the spine switches is to make routing decisions for the overlay. The primary design considerations when choosing the best switch for your data center spine are:

• Port density
• Ports speeds
• Routing table sizes

Leaf Switches

There are two leaf switch models to choose from in the Aruba ESP Data Center reference architecture. Both models are 1RU ToR switches that support high density racks using 10GbE copper or SFP+ ports. SFP ports on the 8360 model also support 10GBASE-T transceivers

For redundant ToR designs, the high and medium density SKUs provide the minimum of four uplink ports that are required for a two spine switch topology. For non-redundant ToR design, medium and low density SKUs provide the minimum of two uplink ports required for a two spine switch topology.

您可以在共同的脊柱和叶片拓扑中混合和匹配不同的TOR配置，以便优化主机聚合需求。下表总结了具有相应支持的设计的叶子SKU。

Out-of-Band Management Switches

The Aruba ESP Data Center reference architecture uses a management LAN built on dedicated switching infrastructure to ensure reliable connectivity to data center infrastructure for automation, orchestration, and traditional management access. The following table lists the recommended switch models.

基于“增大化现实”技术uba Fabric Composer

Aruba Fabric Composer作为独立ISO或虚拟机OVA提供，可以在虚拟和物理主机环境中作为单个实例或作为高可用性，3节点集群安装。Aruba Fabric Composer可作为每次交换机软件订阅作为年度。

基于“增大化现实”技术uba Fabric Composer supports Aruba CX 8325 and 8360 switches.

Aruba Fabric Composer的订购信息在结束时solutions overview.

NetEdit

NetEdit runs as an virtual machine OVA on a host. Aruba NetEdit is available from the Aruba Service Portal. Customers must visit the Aruba Airheads Community and create an Airheads account in order todownload the NetEdit software.

Ordering information for Aruba NetEdit is at the end of thedata sheet.

Reference Architecture Physical Layer Planning

以下部分提供了guidance for physical layer planning of your data center switches.

Cables and Transceivers

Please refer to the following documents to make sure that you select supported cables and transceivers as you plan for physical connectivity inside your data center:

HPE Server Networking Transceiver and Cable Compatibility Matrix

基于“增大化现实”技术ubaOS-Switch and ArubaOS-CX Transceiver Guide

港口速度组

When planning for ToR configurations that require server connectivity at multiple speeds, it is important to note that setting the speed of a port might require adjacent ports to then operate at that same speed.

The CX 8325 series switch has a default speed of 25GbE. To change the speed to 10GbE will impact groups of 12 ports. The CX 8360 series switch allows individual ports to operate at different speeds without impacting adjacent ports unless MACSec is in use. Ports configured to use MACSec must all be configured to operate at the same speed.

Split Ports

Breakout cables can be used to split a 40 Gb/s or 100 Gb/s port, into four lower speed connections (4x10 Gb/s and 4x25 Gb/s). Please refer to the ArubaOS-Switch and ArubaOS-CX Transceiver Guide for selecting supported breakout cables and switch support for split ports.

Media Access Control Security

媒体访问控制安全性（MACSEC）是IEEE 802.1A中定义的标准，它扩展了标准以太网，以在点对点链路上提供帧级加密。该特征通常用于需要其他数据机密性层的环境中或者无法物理地保护系统之间的网络链路。有关Aruba交换组合中的Macsec支持的详细信息，请参阅下表：

Reference Architecture Capacity Planning

The following section provides capacity planning guidance for the Aruba ESP Data Center spine and leaf reference architecture.

带宽计算

A spine and leaf network design provides maximum flexibility and throughput for your Aruba ESP Data Center implementation. To achieve the greatest level of performance, a spine and leaf topology can be designed for zero oversubscription of bandwidth. This results in a data center network that will never be congested because the bandwidth available to hosts is equal to the bandwidth between leaf and spine switches.

A significant advantage of a spine and leaf design is that additional capacity can be added as needed by adding additional spine switches and/or increasing the speed of the uplinks between leaf and spine switches. A rack with 40 dual homed servers with 10GbE NICs could theoretically generate a total load of 800G of traffic. For that server density configuration a 1:1 (non-oversubscribed) fabric could be built with four spine switches using 4x100GbE links on each. In practice most spine and leaf topologies are built between 2.4:1 and 6:1 server to fabric oversubscription ratio.

Network and Compute Scaling

Aruba ESP数据中心参考体系结构为大多数部署提供足够的容量。通过使用分布式网关和对称IRB转发Mac和ARP表将直接连接的计算节点直接连接，并且不会受到部署的机架量的影响。IP前缀的量将是节点，结构链接以及物理和/或虚拟化服务器的数量的函数。边框叶通常是具有最高控制平面负载的节点，因为它处理内部和外部连接。路由摘要是一种良好的做法，可以减少域之间的IP前缀的重新分配。

Aruba数据中心参考架构在结束于端解决方案环境中彻底测试，该环境包含最佳实践部署建议，应用程序和代表生产环境的加载配置文件。

Please refer to the product data sheets on基于“增大化现实”技术uba Campus Core and Aggregation Switches有关未包含在本指南中的详细规格。下表提供了可验证的多维配置文件，可以用于脊柱和叶设计的容量规划。

---