Warakorn - Fotolia
9 VDI storage best practices to better connect remote workers
Learn how to plan and deploy VDI storage to make your remote workers more productive and efficient. It's all about knowing your users and picking the right devices.
Providing storage for VDI is no small task. Workloads fluctuate, users have different requirements and many now work from home. IT teams must balance storage performance against capacity and costs, as well as accommodate necessary data protection capabilities, all of which add to the overall complexity.
To complicate matters, many VDI deployments support persistent desktops, which require more capacity for both the primary data and backups. To help IT deal with all these factors, look at these nine VDI storage best practices for effectively planning and deploying the technology in the enterprise.
1. Understand your users and their work patterns
For a VDI implementation to succeed, the storage system must support the users and how they work. This approach starts with understanding their requirements and work patterns. VDI storage traffic is typically random and fluctuates throughout the day. It often has heavier write loads than reads. Storage drives must be able to deliver the required IOPS no matter the circumstances and still accommodate the necessary capacity, while being able to adapt to future requirements.
IT teams must assess the types of workers they'll be supporting. Are they productivity or knowledge workers? It's also important to know what applications they'll be running -- productivity, video editing or computer-aided design. IT should also determine which guest OS users will need. Although many will be on Windows 10, some might require earlier Windows versions or Linux desktops, which can make a difference in IOPS and capacity. In addition, IT should verify how much data users will store and whether they'll require persistent or nonpersistent desktops. It's also necessary to verify the number of users, as well as when and where they'll work.
2. Think security
As a result of COVID-19, more people are working from home than ever, making security and compliance a greater concern. Although virtual desktops can help protect sensitive data, the growing number of remote workers adds risks, such as people working in less secure environments, adopting informal work habits and using shadow IT. A remote workforce also creates more opportunities for insiders to carry out malicious behavior.
When implementing VDI storage best practices, VDI storage administrators should use granular data protection policies that provide comprehensive identity and access management, applying the principal of least privilege to all resources. They should also encrypt data at rest and in-motion and implement endpoint security protections wherever possible. Safeguarding data in a VDI environment is part of a larger security and compliance strategy that protects data wherever it resides and provides employees with both education and tools for safely working from home.
3. Don't forget the VDI product components
A VDI platform is made up of multiple components, many of which have storage requirements beyond what's needed to support the virtual desktops, applications and user data. For example, Citrix Virtual Apps and Desktops includes Delivery Controller, Studio, StoreFront and License Server, all of which require adequate storage. In addition, the Citrix platform needs storage to support the SQL Server databases for monitoring, site configuration and configuration logging.
A platform's hypervisor also requires storage for every host it runs on, not only for the software itself, but also for any files it creates. For example, the hypervisor might generate log files for each VM running on that host. Although the disk space needed for the product components is typically much less than for the virtual desktops and data, IT must still accommodate their storage requirements, including any optional features and services that might be implemented.
4. Choose the right storage platform
For on-premises VDI deployments, IT will likely deploy NAS or a SAN, rather than DAS, unless it opts for hyper-converged infrastructure. One VDI storage best practice is to select a system based on the size of the VDI environment and the specific requirements. In some cases, an existing storage system might be able to support the VDI implementation, but be careful about trying to use legacy storage that's not up to the job.
The team will also need to determine the best storage protocols and interfaces to use, considering the storage platform, whether it uses file, block or object storage, and which network fabric it uses -- Ethernet, Fibre Channel (FC) or InfiniBand. For example, if an IT team is implementing an all-flash SAN array on a FC network, it might use NVMe-oF and deploy host bus adapters that can support NVMe over FC. However, if it's deploying cloud-hosted VDI, then it will likely use the storage services available through the cloud provider.
5. Choose the right devices
When planning the storage platform and its supporting infrastructure, IT teams should also address the storage devices themselves. Traditional VDI systems have relied on HDD storage, but these are quickly giving way to SSDs. Not only can they deliver substantially greater IOPS at much lower latencies, but they're cheaper, more durable and support greater capacities. In some cases, an IT team might opt for hybrid storage, which uses both HDDs and SSDs. Storage tiering is an option, putting IOPS-intensive workloads on the SSDs and less demanding workloads on HDDs.
When selecting devices, IT should evaluate the built-in software or firmware that manages storage operations and maximizes performance and durability. These features vary among drive types and storage platforms and should be carefully weighed. Many systems incorporate data reduction technologies such as deduplication, compression and zero detection, although their effectiveness depends on how the VDI environment has been configured. For example, data reduction techniques are more effective for full clones, as opposed to linked clones.
6. Choose the right technologies
IT teams must determine which technologies they'll implement to support their storage systems. For instance, if they plan to implement RAID, they should decide which RAID level to use. The right level can improve performance and provide fault tolerance.
Another technology worth considering is software-defined storage, which makes it easier to manage and scale systems. IT teams should also determine how they'll monitor and analyze the storage systems and their workloads to ensure optimal performance and protect data.
Another important storage decision is related to how local caching will be implemented. Even in persistent VDI environments, less expensive storage can be used to offset the price of solid-state drives while still providing the performance VDI needs if caching is used for I/O-intensive periods such as boot storms. Caching should be used as much as possible. Store linked clones on solid-state drives. Even user profile data can be cached to maximize performance.
7. Take advantage of product-specific options
VDI products often provide features for improving storage performance and reducing capacity. One VDI storage best practice is to use cloning capabilities to create a base image and from that, create multiple linked clones that share the base image, substantially reducing the amount of stored data. Another option is thin provisioning to allocate disk space on demand, rather than reserving it all upfront.
VDI products often offer unique features. For example, VMware Horizon provides vSAN for software-defined storage, View Storage Accelerator to reduce IOPS requirements and Virtual Volumes for optimizing SAN and NAS arrays for a virtual environment. And Citrix Virtual Apps and Desktops includes Machine Creation Services and Provisioning Services, which deliver better storage performance and reduce VM storage requirements. IT teams should consider these types of features when evaluating VDI platforms.
8. Accommodate VDI disaster recovery
Not only must VDI storage be able to handle day-to-day operations, but it should support disaster recovery. VDI storage best practices require replicating user data to a secondary site to protect against data loss. Another approach is to set up your database system so it supports the VDI implementation in a failover cluster to provide high availability and guard against disaster. In such cases, the storage systems must accommodate the additional capacity and loads required without affecting primary operations.
VDI often generates backups and snapshots to protect virtual desktop environments and their data. A system will often back up desktop images, as well as support files, such as configuration and database log and data files. Snapshots might also be used to preserve VM state, which can add considerable storage overhead. In addition, VDI can be set up in active-active clusters to provide high availability, incurring even more storage overhead. IT teams must also ensure the storage systems themselves are configured with redundant components such as dual controllers.
9. Plan for remote workers
Now more than ever, IT must consider the desktop needs of at-home workers. Virtual desktops are well suited to remote workers, but they can suffer from high latency and poor performance if the infrastructure isn't architected with location in mind. If people aren't in close proximity to the data center where the VDI platform is hosted and the data is stored, they could have problems accessing their desktops.
If a workforce is highly distributed, it might be wise to deploy multiple VDI sites in different locations and, for each site, provide storage that's close enough for optimal performance. Cloud platforms can be beneficial in supporting this scenario, but even then, distributed VDI requires a way to facilitate redundancy, collaboration and data sharing. For this, an organization might turn to a cloud-based service such as Nasuni's that provides data access across distributed VDI sites.