Give your server a bigger throat

Give your SERVER a bigger throat!

Deliver more data from the same hardware

Increase the efficency of your server

Run more virtual machines

Create a greener environment

Servers are limited by how intelligently they access storage. Normally, a peripheral controller manages the storage accesses for a server, but a peripheral controller is not the best choice here. Peripheral controllers are stuck out on the other side of the I/O bus they plug into, where they're isolated from the operating system with only a limited amount of memory for buffering, limited CPU power for planning what to access, and a complete lack of file system knowledge to guide their decisions.

The result is that servers typically make inefficient use of storage, causing disks to saturate¹ before they need to. Imagine an extreme case where 1 MByte is requested from a disk for an application by issuing 1000 1 KByte requests. Each request will have its own mechanical seek time, pushing the time to get the 1 KByte out to several seconds. One request for the full 1 MByte would get the same data in a few milliseconds, with 999 fewer seeks. This is the kind of intelligence NewServe applies, but on a much more sophisticated level.

By managing storage from inside the operating system, NewServe is able to use the rich resources and information there to pull the same data from the disks with less mechanical motion than is possible with peripheral controllers. NewServe effectively gives your server a bigger throat by allowing clients to pull more data from the disks before saturating them. The benefit to you is that more clients can be served from the same hardware. (Maybe you don't really need that second server you were pricing.)

This is how servers should underline work. The US Patent Office awarded a patent to NewServe in 2010.

1. A disks saturates when there is no longer any idle time between the disk accesses: the disk arms are constantly seeking from one location to the next, with no free time between the accesses. At this point, sending more request streams to the disk does not increase the amount of data returned by the disk and the disk is said to be saturated. The situation is like a bank: when all the tellers are working as fast as they can, sending more customers into the bank doesn't increase the amount of money transferred by the bank.