Kirill Kedrinski - Fotolia

Tip

Don't let VDI network latency grind virtual desktops to a halt

Network latency problems can make even the simplest tasks impossible on a virtual desktop. IT needs to use performance monitoring to identify the source of latency to fix the problem.

VDI network latency can cause huge problems for virtual desktop users. IT needs to know the best ways to protect its user experience, or else suffer the fallout.

In extreme cases, VDI latency problems inhibit users' ability to type, because each keystroke can take several seconds to appear on their virtual desktops. Even in less extreme situations, latency has a direct effect on the end user experience, and widespread dissatisfaction can spell doom for a VDI deployment.

Choose the right battles

The number one rule for network administrators dealing with latency is to focus on what they can control. There are countless causes of latency problems that network admins simply have no power over. Imagine, for instance, a user works on a laptop in the break room. Everything works fine for the user until someone decides to microwave a burrito. The microwave disrupts the Wi-Fi signal, which results in packet loss and induces latency. There's not much IT can do about that.

The only way to effectively deal with VDI latency is to address the underlying cause.

Latency can also be an issue for users who connect to virtual desktops from public networks. The user's internet connection may be slow, which results in poor VDI performance, even though the VDI setup itself is healthy. Admins can help themselves out by communicating to users when a latency issue is out of their control, or by educating workers on best practices to ensure strong network connectivity. IT should primarily focus, though, on ways it can optimize the performance of users' virtual desktops.

Remember, it is impossible to fully control network latency problems, but there are some guidelines administrators can follow to ensure the best possible VDI user experience.

How IT can fight latency

First, IT needs to define what it considers a normal amount of latency in its VDI deployment. That barometer is going to vary considerably from one organization to the next, but VDI vendors typically provide network bandwidth recommendations. VMware, for example, suggests that IT plan for 100 to 150 Kbps for basic Horizon View users who do not work with video or 3D graphics.

After establishing an acceptable amount of latency, it is important to monitor the network's performance. Performance monitoring tools are useful for troubleshooting problems, but also for spotting long-term trends. Whatever performance metrics IT records initially can serve as a benchmark to compare against any future metrics. That way, an administrator who notices a gradual increase in latency over time can add additional network capacity before the latency causes problems.

Network performance monitoring also helps IT determine the source of latency. The only way to effectively deal with VDI network latency is to address the underlying cause, so IT needs to know where the bottleneck originates. The latency problems may be tied to one specific network segment or to the storage infrastructure. Regardless of the latency's source, performance monitoring helps admins find the origin of the problem, and it provides documented proof of the issue in case they have to ask management for funding to fix the issue.

Take advantage of existing resources

Another key to avoiding VDI latency issues is to use existing data center resources more efficiently. For example, IT could implement a Quality of Service (QoS) tool to help ensure a good VDI user experience. Admins can use a QoS tool reserve network bandwidth for a particular purpose, such as for Microsoft Remote Desktop Protocol traffic. Although a QoS tool won't increase the overall network bandwidth availability, IT can force the network to give preferential treatment to the traffic that matters most.

Admins may also be able to reduce VDI network latency by moving certain traffic types, such as cluster communications or replication traffic, to dedicated network segments. IT can also use network interface card teaming to increase the network bandwidth available to its VDI servers.

IT should also see if it can use any features in its virtualization software to improve user experience. For example, Citrix XenApp and XenDesktop include a feature called SpeedScreen Latency Reduction (SLR), which creates a local echo of typed characters before transmitting the keystrokes to the server. This makes users' keyboard and mouse instantly responsive within their virtual desktops. The SLR feature also uses an hourglass pointer to show when the server has not yet registered a mouse click. This helps prevent users from clicking the mouse multiple times, resulting in unpredictable behavior.

Solving VDI latency issues is rarely a simple process, but there are many tools available to help IT address these problems. The first step must be to embrace network performance monitoring, because determining the source of a latency problem is the only way to permanently fix it.

Next Steps

Don't let demand spikes cripple VDI user experience

Storage infrastructure's role in VDI performance

How to avoid VDI boot storms

Dig Deeper on Virtual and remote desktop strategies