Most, if not all of the cloud, was intended to be centralized for the delivery of services such as applications, desktops, and data; however, the network has been an understood bottleneck since the beginning. Should anyone achieve centralizing their data, the complications and performance degradation, compared with local models of delivering those services, are barriers to adoption at best, but quite often prevent the adoption of the model from moving forward all together.
It has been said “…the one which is closest to the current PC Lifecycle Management will be the one which succeeds where others fail”. The same holds true for components of the PC such as applications and data. In the current model data and applications are on every device, distributed everywhere. Desktops have found some adoption in the cloud as an extra desktop or for shift worker type use cases but cloud based desktops for knowledge workers or specialized workloads have known issues of performance, availability and other disadvantages compared to their local equivalents.
Quite simply, distributed architectures overcome bottlenecks by bringing the ideal workloads closest to the user who needs and uses them…while connecting them to hosted environments for shared information and resources such as data analytics, on-demand upgrades, and failover resources. This conclusion has been considered just an idealistic way of thinking because we lacked the tools to approach things in a distributed way… Until now!