juanjo tugores - Fotolia
One of the driving forces behind the desktop as a service trend is data gravity. As data centers grow to hold more applications and data, it only makes sense that desktops are attracted to the cloud, too.
The concept of data gravity is that large amounts of data tend to attract more data. As data centers grow to hold more data and applications, they generate more gravity. And as cloud data centers attract more corporate data, they will also attract desktops. As long as you have client/server applications, data gravity is going to attract the client to the server.
A central cause of this is that data is more valuable if users can access it quickly. Data that is physically closer to users equals quicker access, and related data tends to be stored in the same location give the greatest value. Desktop as a service (DaaS) simplifies the movement of the desktop into the cloud alongside the servers.
Data gravity decreases latency
One of the fundamental limits in life is the speed of light. For computers, it means the farther apart two computers are from one another, the slower they will communicate, also known as latency. Application performance relies on the combination of PC power, server power and the speed of sending the data.
Modern Web applications are designed to do most processing in the data center and rely on a Web browser to display the results, but older client/server applications -- some of which are mission-critical -- expect to split data processing transfer between the data center and users' PCs.
PC and server power have vastly increased every year for the last fifty years, as has the speed of the local network. And developers have written a lot of software with the assumption that the networks are fast and low-latency. But when you move an application from a server in your building to a cloud provider's server in another state -- or on the other end of the country -- you change the distance between your users' PCs and the servers. For some applications, this change can lead to disastrous performance.
One way to reduce data latency is to put users' desktops in the same data center as the applications they need. You can use remote display protocols such as HDX and PCoIP to deliver the desktops -- these protocols cope with latency much better than client/server applications do.
What does DaaS got to do with it?
Before DaaS, hosting desktops in the cloud meant deploying your own virtual infrastructure or Remote Desktop Session Host into that cloud alongside your data. Adding VDI platform management could eat up all of the cost savings that you expected from using the cloud to host data and applications in the first place.
The desktops live inside the provider's data center, so they are close to your servers and therefore close to your data and applications. You can rent a few or a lot of desktop VMs, and there are a lot of different DaaS platforms and providers with competitive options and price points to choose from. The provider manages the entire infrastructure required for VDI, and shares the cost across multiple customers. The provider can also spread the cost of the expertise required to build a good VDI deployment.
Most organizations need client/server applications to operate, and until these applications are replaced, we need to keep the desktop application close to the server. If you move the server into a cloud provider's data center, that same data center will attract users' desktops. Star Trek's Scotty the Engineer said it best: "Ye cannae change the laws o' physics."
Network latency effects on application performance
Eliminating VDI network bottlenecks and latency issues