Running Avaya ERS with VMWare ESX Server
No doubt about it – you ARE virtualizing something within your data centre…Perhaps the majority of your applications are already virtualized or you are just beginning to consolidate your servers now. Either way a hot industry trend for saving money and utilizing resources to the full potential.
Consolidating is great to save money and resources, however ensuring corporate up time is still no doubt a number 1 priority! That’s what you invest in server architecture to support that up-time and resiliency!
But what about your data architecture?
Business Up-time Policy
Both the server hardware and the networking topology must eliminate any single points of failure, provide instant fail-over capability for link or node failures, and ideally maximize the use of deployed
Both the server hardware and the networking topology must eliminate any single points of failure, provide instant fail-over capability for link or node failures, and ideally maximize the use of deployed switches, network interface cards (NICs), and cabling.
switches, network interface cards (NICs), and cabling. Those using Avaya ERS 8800, 8600, 8300, and ERS 5500 architecture today will be familiar with Avaya (Nortel) Switch Clustering as a solution that delivers unmatched resiliency, simplicity and cost effecting performance.
Couple the Avaya Switch Clustering architecture with Avaya’s VMware best practices for virtual networking and you then provide your business with the “server AND Networking topology eliminating single points of failure” that you and your business depend on.
Avaya provides a document for best practices Resilient Data Center Server Edge Solutions for VMware ESX Server Technical configuration Guide and you can find it located further down in this post. This article provide a quick overview of the document. For further info and specific configuration details and commands, please download the Technical Document
Accomplishing the Resiliency
Afew recommendations from the VMware aspect as well as Avaya’s are provided here to take into consideration when deploy Avaya’s best practice.
VMware Virtual Networking Recommendations Include:
- Separate Service Concole/VMotion/ISCSI from VM Traffic
- Cross-Team On-Board NICs with PCI NICs
- Use Virtual Switch Tagging for VLAN assignment and isolation
- Disable Spanning Tree Protocol
Nortel Recommendation for connecting VMware ESX include:
- Use ERS Switch Clustering (SMLT) to the server NIC level
- Use Horizontal Stacking and Switch Clustering to simplify cabling and provide zero
service impact maintenance
- Use ESX NIC Teams in IP Hashing Mode to balance traffic across NIC team *and* take
advantage of multiple switch redundancy
No doubt about it, to provide resilient stacking you gotta be using SMLT functionality. See this GUIDE for more details on configuring SMLTs
A FEW OPTIONS TO CONSIDER
1. Single Server Resiliency
This single server solution provides high-density Gig-E edge switching for servers running ESX as well as multi-switch resiliency.
2. Advanced Multi Switch Server Resiliency
Solution provides high-density Gigabit Ethernet edge switching for ESX servers and provides
multi-switch resiliency with Nortel Switch Clustering.
3. Advanced Multi Rack (Horizontal Stack) Switch Server Resiliency
Moving further into denser server environments with multiple server racks. The practice of providing Horizontal Switching Server Resiliency comes into play. This is provided by combining both Stacking and Switch Clustering with SMLTs. Enabling both of these configurations will give you the ability to connect large numbers of servers with dense in-rack Gigabit connections with active/active links and all of the resiliency benefits of clustering.
No complex configuration is provided as this solution is a simple extension of the topology shown in the above section “Singer Server Resiliency”
>Note that each unit that is added to a stack adds 80 Gbps of aggregate bandwidth to the virtual switching fabric providing unmatched throughputfor access to uplinks and inter-stack switching.
Implementing a Horizontal stacking design will use a unique application of the stack cables to connect multiple top of row ERS 5500 switches. This will extend the stack across a number of server racks or cabinets as shown
Horizontal Stacking Provides the following Benefits
As stated by Avaya
- Fault-tolerant or Load sharing NIC teaming into stack
- Low latency between servers (9μs)
- Highly resilient stacking technology with scalable uplinks
- The ability to leverage two or more 10GbE uplinks across the stack
- Flexibility to spread across multiple data cabinets (100s of servers)
- Ideal for Grid Computing / High-Performance Computing Solutions
- Very High Node-to-Node Communications Bandwidth without impacting uplink capacity:
Ideal for VMotion and iSCSI
For full information, technical details, and how to configure any of the above mentioned scenarios. Please feel free download this document.
Let me know if this helps, and what you are deploying with your architecture. I just got emailed a question on Avaya configuration guide on supporting Hyper-V. I will update this post shorltly with information on Hyper-V!