Virtualizing Microsoft Exchange is one of the primary use cases that we see for HC3 customers. A general move to virtualizing Exchange has gained traction as companies take the normal cycle of hardware refreshes and Operating System upgrades as an opportunity to consolidate servers in a virtualized environment. These companies seek to take advantage of:
Our customers rely on HC3 to provide the level of availability and performance required given the mission-critical nature of the application. In this post, I’ll be covering the HC3 features most utilized in an Exchange deployment.
The simplicity of HC3 is that all VMs created on the system are automatically made highly available. In the event of a node failure, the VMs running on the failed node will automatically failover to other nodes in the system with no manual intervention from the user. HC3 VM Failover™ relies on the HC3 Protect™ feature that effectively acts as network RAID 10, by striping and mirroring data across the cluster so that it can tolerate any single point of failure. Because of HC3 Protect™, disk failures in the HC3 environment also have little effect on the VMs running on the system. Following a disk failure, HC3 will automatically re-generate mirror copies of any data blocks that had been stored on the failed disk which will result in a temporary increase in read and write load on the remaining disks in the cluster. If a full node is off-line, its missing disks would reduce the overall number of disks in the cluster available to process I/O.
Often we see users deploy multiple role server VMs when using Exchange for load balancing across the cluster. HC3 allows users to live-migrate these VMs to redistribute the workload across the cluster non-disruptively using the Move feature. This will keep workloads up and running while the host is changed from one node to another in the system.
As the workloads in a user’s environment grow, adding compute and storage capacity to an HC3 system is as simple as adding a node to the rack and giving it an IP address. The system will subsume the resources of the new node into the system without any downtime adding to both the capacity and the performance of the environment. Once new nodes are added, users can create new role servers such as an Exchange Mailbox Server to expand the capacity of Exchange deployment. This process is magnitudes faster than the procurement and setup of a new server in a physical infrastructure and requires no downtime.
The HC3 storage layer uses “wide striping” to distribute I/O load across all disks in the cluster, aggregating their performance as well as providing the data storage redundancy described in the HC3 Protect™ feature above. Adding additional nodes to the cluster adds more disk IOPs improving the overall performance of Exchange when running on HC3. This allows a wide range of demanding mailbox servers to perform well on HC3 systems as demonstrated in our Microsoft Exchange Solution Reference Architecture Application Note. In addition to the IOPs required for Exchange, HC3 VM’s support up to 8 vCPU’s up to 60GB RAM per VM (on certain nodes, greater than 30GB not recommended to optimize VM failover placement), which is significantly greater than what most Exchange environments require.
A major benefit of the scale-out nature of HC3 is the flexibility that comes when sizing an HC3 system to meet the performance and capacity requirements of the user’s workload. Start with 3 like-sized nodes and then mix and match nodes with differing drive capacities and speeds as appropriate to scale-out the system as needed. With the variety of HC3 systems that can be built, Microsoft Exchange on HC3 meets or exceeds the performance and capacity requirements of most small and mid-size exchange deployments. This functionality also allows you to purchase just what is needed for the Exchange infrastructure today and scale the mailboxes and services to support increasing workloads by quickly adding VMs and nodes when the user needs to expand the compute or storage capacity. Multiple exchange roles can be run in separate VMs and load balanced across the HC3 cluster to maximize the performance of the deployment. Scale recommends following Microsoft Best Practices using Microsoft tools to determine how to best split your exchange server roles across the different virtual machines on HC3.
If you are looking for a virtualization solution to provide high availability, simplicity and scalability for your virtualized Exchange workload give us a call! We would love to hear from you.
Please complete the form below and someone from Scale Computing will be in contact with you within 24hrs.×