operating system - Why isn't every OS real-time? -

I have not heard anything about RTOS, but they give the programmer more control over the scheduler so that the priority is inversely Avoid, their time is more consistent, better multitasking but all standard desktop setup uses OS which are not real time. There should be some tradeoffs for using RTOS, what are they?

RTOS typically features general definition of "real-time" people for business performance and forecasting and traceability "Deterministic"; You can not make a fate without paying for it.

In the general purpose OS, we are inspired by "common-case" behaviors - we really want very good performance and much more flexibility. At RTOs, we want a reliable roof over the "worst condition" behavior, and we pay in throughput or normal-type behavior (often referred to as). Yes, it is possible to make hybrids, such as Windows or Linux real-time threads but somewhere you are usually finishing because only a limited set of available resources (CPU, IO bandwidth, whatever) and consumer Different standards of OS and RTOS are compatible. Some RT-Linux approaches clearly deal with this, by division

What is the business turnover? I can not give an exact list - this is more general purpose OS. S is a district engineer, and is able to maintain the churn of new equipment; RTOS focuses on a very small set, for which timeliness can be well understood or intercepted with interference with other activities. You will not have the same selection of drivers on normal RTOS because it is not appropriate for them to apply, usually

Throughput Remember "real-time"! = "Real-fast" when a system is in real time, it means that the time of completion activities is part of their accuracy. In some cases, this means that processing many activities very quickly (high throughput) ; In others it may be processing on a relatively slow but very predictable period. The structure can be high throughput in an RTOS, but usually does not get the equivalent of an RTOS, because the techniques (caching, fancy interactivity-driven scheduling approaches, "fair" queue And lock dispute) against the prediction of timeliness of a single work.