Partitioning Using vPars

To understand how vPars works, compare it to a server not using vPars. Figure 2-1 shows a 4-way HP-UX server. Without vPars, all hardware resources are dedicated to one instance of HP-UX and the applications that are running on this one instance.

Figure 2-1 Server without vPars

Figure 2-2 shows the software stack where all applications run on top of the single OS instance:

Figure 2-2 Software Stack of Server without vPars

Using vPars, you can allocate a server’s resources into two or more virtual partitions, each with a subset of the hardware. In Figure 2-3, two virtual partitions are shown, each with its own boot disk, its own processor resources, its own LAN connection, and a sufficient subset of memory to run HP-UX and the applications intended to be hosted on that virtual partition.

Figure 2-3 Server Block Diagram with Two Virtual Partitions

Each application can run on top of a separate OS instance. Instead of a single OS instance owning all the hardware, the vPars Monitor manages the virtual partitions and each OS instance as well as the assignment of hardware resources to each virtual partition.

Figure 2-4 Software Stack for Server with Two Virtual Partitions