How It Works
Parabon Computation Grid
A New Computing Paradigm: Pay-per-use Computation versus Pay-whether-you-use Computers
A schematic view of the Frontier Grid Platform
Launched in 2000, the Parabon Computation Grid is the longest continuously running commercial grid in existence. Managed by our Frontier Grid Platform software, the Parabon Computation Grid is the computation utility infrastructure, comprised of servers and compute nodes worldwide, that provides the power behind our Frontier Grid Services.
- No Hardware SM. The Parabon Computation Grid obviates expensive, hardware-centric "brick-and-mortar" sources of computation with an online, on-demand service model that fundamentally changes previous notions of computational capacity and how it is purchased.
- Buying time. This utility service is unique because it is powered, not by dedicated servers in a data center, but by the excess computational capacity we purchase from our institutional providers (mostly businesses and universities) on the Parabon Capacity Market.
- Waste nothing. Because of the "peaky" nature in which computers are used, excess capacity is plentiful (estimates put server and desktop utilization rates at just 5-20%). Accordingly, we can offer high-performance computation far more affordably than utilities with dedicated resources, which have high overhead costs.1
- Think big. Moreover, Frontier's practically limitless scalability — we can always contract for more capacity when demand warrants — means our computation utility can service larger jobs than any other utility on the planet.
- Go green. Lastly, it is environmentally friendly. By making use of otherwise wasted capacity, Frontier helps offset the enormous carbon emissions generated the massive data centers that power some computational services.
Differences between Hardware and No HardwareSM approaches to computational demand
| Hardware | No Hardware™ |
|---|---|
| Clusters, server-farms, supercomputers | Parabon's Computation on Demand™ service |
| Fixed and limited capacity | Variable and extreme-scale capacity |
| Underutilization from peaky demand | Full utilization that exactly matches demand |
| Capital expenditure that hyper-depreciates | Expense item that increases in value |
1 Taurus - A Taxonomy of the Actual Utilization of Real UNIX and Windows Servers, David G Heap, Principal IT Consultant, IBM Enterprise Server Group, January 2003.
