Руководство По Проектированию для Cisco Cisco Nexus 5010 Switch
Design Guide
© 2010 Cisco Systems, Inc. All rights reserved. This document is Cisco Public Information.
Page 20 of 22
●
The ability to look at the interface counters for the virtual network adapter connecting the virtual machine to
the VEM. These counters are the same ones available on any Cisco switch for physical interfaces, and they
are statefully maintained in case a virtual machine moves from an ESX host to a different one.
the VEM. These counters are the same ones available on any Cisco switch for physical interfaces, and they
are statefully maintained in case a virtual machine moves from an ESX host to a different one.
●
The ability to use an Encapsulated Remote Switch Port Analyzer (ERSPAN) with a virtual Ethernet interface,
even if the virtual machine migrates.
even if the virtual machine migrates.
●
The ability to apply security policies, such as access lists, private VLANS, and port security, on virtual Ethernet
interfaces.
interfaces.
The Cisco Nexus 1000V Switch performs all these operations and Layer 2 forwarding without introducing any Layer 2
loops. It does so without the need to run the Spanning Tree Protocol on the server.
loops. It does so without the need to run the Spanning Tree Protocol on the server.
Chapter 8 provides more details on how to design with the Cisco Nexus 1000V Series Switches on virtualized
servers.
servers.
Software Revisions
The testing that supports this design guide was run on a topology consisting of:
●
An aggregation layer of the Cisco Nexus 7000 Series Switches running Cisco NX-OS Software Release 4.1(5)
●
An access layer of the Cisco Nexus 5000 Series Switches and Cisco Nexus 2000 Series Fabric Extenders
running NX-OS Software Release 4.1(3)N1(1)
running NX-OS Software Release 4.1(3)N1(1)
●
Virtualized servers with Cisco Nexus 1000V Series Switches running Cisco NX-OS Software Release
4.0(4)SV1(1)
4.0(4)SV1(1)
These code revisions do not constitute a code recommendation. New code versions add valuable enhancements and
new features.
new features.
Core, Aggregation, and Access Layers
Data centers based on Cisco Nexus products follow the well-proven Cisco architecture of core, aggregation, and
access layers. The details of the functions provided by each layer can be found in the Cisco white paper entitled Data
Center Design—IP Network Infrastructure
(
access layers. The details of the functions provided by each layer can be found in the Cisco white paper entitled Data
Center Design—IP Network Infrastructure
(
) In addition,
design guidelines for the aggregation layer based on vPC can be found in Chapter 5, which guidelines for the access
layer can be found in Chapter 6.
layer can be found in Chapter 6.
As a quick-summary reference, the aggregation and access layers provide the following functions:
●
The aggregation layer of the data center provides a consolidation point where access layer switches are
connected. This, in turn, provides connectivity between servers for multitier applications, as well as
connectivity across the core of the network to clients residing within the campus, the WAN, or the Internet.
The aggregation layer typically provides the boundary between Layer 3 routed links and Layer 2 Ethernet
broadcast domains in the data center.
connected. This, in turn, provides connectivity between servers for multitier applications, as well as
connectivity across the core of the network to clients residing within the campus, the WAN, or the Internet.
The aggregation layer typically provides the boundary between Layer 3 routed links and Layer 2 Ethernet
broadcast domains in the data center.
●
The access layer of the network provides connectivity for server farm end nodes residing in the data center.
Traditional data center access layer designs are strongly influenced by the need to locate switches in a way
that most conveniently provides cabling connectivity. The most commonly used designs for data center server
farm connectivity are end-of-row and top-of-rack
Traditional data center access layer designs are strongly influenced by the need to locate switches in a way
that most conveniently provides cabling connectivity. The most commonly used designs for data center server
farm connectivity are end-of-row and top-of-rack
With the widespread adoption of blade servers, the definition of access layer is blurred, as the blade enclosure may
often include a switching component whose main role is either to provide blade-to-blade switching support or simply
to funnel the traffic upstream to the “real” access layer.
often include a switching component whose main role is either to provide blade-to-blade switching support or simply
to funnel the traffic upstream to the “real” access layer.