carloscastilla - Fotolia
Many factors go into crafting a comprehensive and holistic design for VMware environments. VMware Design advice can sometimes be confusing or contradictory, however. With a handful of key considerations, you can ensure that you construct a solid design -- from storage to security to resource management -- to keep your environment running smoothly.
Start by planning for management clusters and considering the balance you must strike with a host-to-VM ratio in production clusters. Then, consider storage, CPUs and security. Finally, keep simplicity in mind at all times and ensure you craft a solid disaster recovery plan, and don't forget to document your process.
Ensure you have a solid design foundation
Key elements to consider when crafting a solid design for your VMware environment include planning for management clusters, right-sizing your hosts and keeping clear documentation about your process.
Management clusters live outside the main production infrastructure. These hosts generally contain critical management tools and applications, such as vCenter, Active Directory controllers, a backup domain name system and Dynamic Host Configuration Protocol servers. Keeping such hosts separate from the main production network ensures network issues can't affect it, so you can more easily get your environment back online in the event of a failure.
Striking a balance between VM density and number of hosts is also crucial to creating a good design. With larger hosts crammed with more VMs, you amplify the effects of an outage and increase the time it takes to restart. However, separating workloads can also increase the effect of host failures, and require you to manage more resources.
Have storage in mind
To build a fully functional VMware design, consider which storage and CPUs you select. CPU core count and speed determine what kinds of workloads will run best in your environment. Transaction-based workloads, such as batch processing, require more cores per CPU socket to process more threads simultaneously, whereas concentrated workloads such as databases favor CPUs with higher speeds and fewer cores.
Factor both local and shared storage into your design. Local storage -- such as spinning disks, solid-state drives and non-volatile memory express cards -- can provide better speed, performance and security; shared storage can help cut down on money and preserve storage space.
Keep your vSphere design simple
A simple approach to vSphere design with reduced complexity ensures users of your environment don't struggle with organizing or navigating many moving parts. More complex elements also add client confusion, increase infrastructure management needs and make troubleshooting more difficult, which means the best design practice is sometimes picking the most obvious choice or designing to meet the bare minimum of your organization's needs.
Only add complexity to a VMware design when clearly beneficial -- such as adding a hyper-converged platform and hiding it from the user. Hyper-converged platforms self-manage their own complexity, so despite the many moving parts, they ultimately lead to simpler management for the rest of an environment.
Secure your environment
To keep your environment safe and secure, craft a security plan into your VMware design. Proper security for a VMware environment should include core VMware offerings, such as NSX, supplemented with additional security offerings, such as AppDefense.
NSX provides security through the process of microsegmentation -- breaking down VMs into logically isolated groups. This means the networking software places firewalls around individual workloads as well as an entire network, so if a specific workload becomes infected with malware or is otherwise compromised, it cannot spread to compromise other workloads in the network. NSX extends this security across the vCenter server and hardware to ensure consistent security policies across an entire data center.
AppDefense adds to this by monitoring and identifying abnormal application behavior and alerting admins to problems. It also has the ability to automatically remediate problems when VM behavior deviates from its intended state. Remediation steps include quarantining, powering off or deleting a problem VM.
Have a disaster recovery plan -- just in case
Failure happens even in the best-designed environments. Knowing what steps to take in the event of a disaster means you can more easily recover from such failures and keep your environment running smoothly.
Begin by assessing your requirements and goals for DR by reviewing your infrastructure and determining what resources are available, and figure out which VMs to prioritize. Then you can come up with a DR plan, which should include a comprehensive prioritization of VMs with each VM having its own recovery point objectives and recovery time objectives. You should also incorporate compute and storage targets -- such as local, remote and cloud storage -- in your DR plan. Make sure you consider workflow in your DR plan, as well.
Once you've come up with a suitable DR plan, ensure you continue to review, update and manage the plan periodically to support changes to your business or data center.