Hard/Soft realtime systems

NAVIGATION

CATEGORIES

Research

REFERRENCE

If you have some good idea, Cilck and mail to us please...

LINKS

Home » Research

Hard/Soft realtime systems
12 answers - 660 bytes -
Hi,
I read that systems are classified as hard realtime and soft realtime
systems. As I _understand_ hard realtime systems are those that are
said to have failed if even a single event does not occur at the right
time. However in a soft realtime system, if an event does not occur at
the time it must, its termed a degradation in performance, not a system
failure.
Now I cannot appreciate this or understand it because it seems like a
subjective classification to me. What is one person's system failure is
only system degradation to another.
Thanks for replying, in case you do !
Really
No.1 | | 881 bytes | |
In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
Really <hk.12AM@gmail.comwrites
>Hi,
>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.
>
>Thanks for replying, in case you do !
>
>Really
>

Is this home work?
No.2 | | 918 bytes | |
"Really" <hk.12AM@gmail.comwrites:

I read that systems are classified as hard realtime and soft realtime
systems. As I _understand_ hard realtime systems are those that are
said to have failed if even a single event does not occur at the right
time. However in a soft realtime system, if an event does not occur at
the time it must, its termed a degradation in performance, not a system
failure.

Now I cannot appreciate this or understand it because it seems like a
subjective classification to me. What is one person's system failure is
only system degradation to another.

Quite true. In addition, how quickly the response is produced
is a factor.

There has been a recent debate in rec.aviation.ifr as to whether
a five-minute lag in ground-based weather radar images getting
to the end-user is quickly enough to be considered realtime.
No.3 | | 958 bytes | |
Really wrote:

Hi,

I read that systems are classified as hard realtime and soft realtime
systems.

'Can be' classified as hard or soft. There are many ways to classify real
time systems. Hard and soft is just one.

As I _understand_ hard realtime systems are those that are
said to have failed if even a single event does not occur at the right
time. However in a soft realtime system, if an event does not occur at
the time it must, its termed a degradation in performance, not a system
failure.

Indeed. It is all a matter of consequences. If your word processor prints a
little slowly nothing of consequence is likely to occur as a result. If a
nuclear reactor system closes a valve to late the consequences could be
truly disastrous. At these extremes it is obvious which is soft and which
is hard. In between is a huge grey area where opinions will differ.

Ian
No.4 | | 1219 bytes | |
Everett M. Greene wrote:

"Really" <hk.12AM@gmail.comwrites:
>>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.
>
Quite true. In addition, how quickly the response is produced
is a factor.

There has been a recent debate in rec.aviation.ifr as to whether
a five-minute lag in ground-based weather radar images getting
to the end-user is quickly enough to be considered realtime.

Speed is irrelevant - what is essential is whether or not a deadline is met
whether it be nanoseconds or hours.

Ian
No.5 | | 1895 bytes | |
In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
Really <hk.12AM@gmail.comwrote:
>Hi,
>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.

The most important thing to remember is that realtime is not the same as
real fast. The deadline might be measured in hours.

My favorite hard realtime requirements happen in industrial automation
and aviation. Have you ever seen a high-speed mail sorter? Scanning
the envelope and then commanding the sorting gate is "soft realtime",
in that if the mail goes into the wrong chute, it's not fatal.
the other hand, on that same sorter, if the sorting gate is in an
intermediate position when an envelope hits, you've just destroyed
several hundred pieces of mail, and possibly damaged the equipment.
That's hard realtime.

aviation, which is a nice show of determinism's need The system
that controls the thrust reversers on landing had better get the timing
right. You don't want them deploying when you're 50 feet above the
ground, and you don't want them deploying after you're run through
the blast fence at the other end of the runway because the deflectors
decided to take 10 seconds to deploy because a GC was needed.
No.6 | | 1575 bytes | |
Ian Bell <ruffrecords@yahoo.comwrites:
Everett M. Greene wrote:
"Really" <hk.12AM@gmail.comwrites:
>>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.
>
Quite true. In addition, how quickly the response is produced
is a factor.

There has been a recent debate in rec.aviation.ifr as to whether
a five-minute lag in ground-based weather radar images getting
to the end-user is quickly enough to be considered realtime.

Speed is irrelevant - what is essential is whether or not a
deadline is met whether it be nanoseconds or hours.

Speed is very relevant. If you set the time limit to infinity,
all computing will meet a hard realtime "requirement" (barring
failure to terminate). Results have to be available in a timely
enough manner to be useful to be considered realtime. That's
the whole point of the debate in rec.aviation.ifr.
No.7 | | 383 bytes | |
Everett M. Greene wrote:

>>
>Speed is irrelevant - what is essential is whether or not a
>deadline is met whether it be nanoseconds or hours.
>
Speed is very relevant. If you set the time limit to infinity,

Precisely, the time limit is important not the speed.Ian

Ian
No.8 | | 945 bytes | |
Nope, its not.
Really

Chris Hills wrote:
In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
Really <hk.12AM@gmail.comwrites
>Hi,
>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.
>
>Thanks for replying, in case you do !
>
>Really
>
>
Is this home work?
--
No.9 | | 1898 bytes | |
In article <1150603843.779236.182350@i40g2000cwc.googlegroups. com>,
Really <hk.12AM@gmail.comwrites
>Nope, its not.
>Really
>
>Chris Hills wrote:
>In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
>Really <hk.12AM@gmail.comwrites
>>Hi,
>>
>>I read that systems are classified as hard realtime and soft realtime
>>systems. As I _understand_ hard realtime systems are those that are
>>said to have failed if even a single event does not occur at the right
>>time. However in a soft realtime system, if an event does not occur at
>>the time it must, its termed a degradation in performance, not a system
>>failure.
>>
>>Now I cannot appreciate this or understand it because it seems like a
>>subjective classification to me. What is one person's system failure is
>>only system degradation to another.
>>
>>Thanks for replying, in case you do !

Hard real time engine management system. Anything misses a deadline
the engine misbehaves and the vehicle crashes.

soft real time airline ticketing system. Used in real time but it does
not matter in reality if it takes 2 or 5 seconds to respond to a
request.

As the airline ticketing system gets business it's performance may
degrade but it is not vital.

Many systems have serious problems if they do not react in time. IT may
be milliseconds or minutes but if it is late it causes serious
problems. E.G. control systems.

whilst many systems are clearly hard or soft it is all subjective around
the edges.

You have to look at the consequences of missing a deadline. This is why
there are 4 SIL's for this sort of work.
No.10 | | 815 bytes | |
Steve Watt wrote:
In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
Really <hk.12AM@gmail.comwrote:
>Hi,
>
>I read that systems are classified as hard realtime and soft realtime
>systems.
>
The most important thing to remember is that realtime is not the same as
real fast. The deadline might be measured in hours.

My favorite hard realtime requirements happen in industrial automation
and aviation.
Steve Watt KD6GGD PP-ASEL-IA ICBM: 121W 56' 57.8" / 37N 20' 14.9"
Internet: steve @ Watt.CM Whois: SW32-ARIN
Free time? There's no such thing. It just comes in varying prices

Hi,
your reply is very informative with live examples of your aviation
domain.
No.11 | | 2081 bytes | |
if missing the dead line means an error, that is hard real time
if missing the dead line just means an unpleasant experience, that is
soft real time

Steve Watt wrote:
In article <1150451105.440573.128640@h76g2000cwa.googlegroups. com>,
Really <hk.12AM@gmail.comwrote:
>Hi,
>
>I read that systems are classified as hard realtime and soft realtime
>systems. As I _understand_ hard realtime systems are those that are
>said to have failed if even a single event does not occur at the right
>time. However in a soft realtime system, if an event does not occur at
>the time it must, its termed a degradation in performance, not a system
>failure.
>
>Now I cannot appreciate this or understand it because it seems like a
>subjective classification to me. What is one person's system failure is
>only system degradation to another.
>
The most important thing to remember is that realtime is not the same as
real fast. The deadline might be measured in hours.

My favorite hard realtime requirements happen in industrial automation
and aviation. Have you ever seen a high-speed mail sorter? Scanning
the envelope and then commanding the sorting gate is "soft realtime",
in that if the mail goes into the wrong chute, it's not fatal.
the other hand, on that same sorter, if the sorting gate is in an
intermediate position when an envelope hits, you've just destroyed
several hundred pieces of mail, and possibly damaged the equipment.
That's hard realtime.

aviation, which is a nice show of determinism's need The system
that controls the thrust reversers on landing had better get the timing
right. You don't want them deploying when you're 50 feet above the
ground, and you don't want them deploying after you're run through
the blast fence at the other end of the runway because the deflectors
decided to take 10 seconds to deploy because a GC was needed.
No.12 | | 1464 bytes | |
Really,

You are right, it is subjective and depends on application. But, if
your medical device inejects more radioactive medicine, becuase it
failed to receive an event, it will be disaster. This is not subjective
to person but a question of life.

Think about an aeroplane, if it explodes because of your system cannot
receive an event on pressure failures in aeroplane, it will not be
subjective, but lost in the air.

My understanding is it depends on the application than a person.
Keyboard input is still real-time, but it is soft. Keyboard response on
telnet sesson over a longdistance telephone cable is still soft, but
less softer than first, as people get panic when they did not see the
key what they pressed.

Ranga

Really wrote:
Hi,

I read that systems are classified as hard realtime and soft realtime
systems. As I _understand_ hard realtime systems are those that are
said to have failed if even a single event does not occur at the right
time. However in a soft realtime system, if an event does not occur at
the time it must, its termed a degradation in performance, not a system
failure.

Now I cannot appreciate this or understand it because it seems like a
subjective classification to me. What is one person's system failure is
only system degradation to another.

Thanks for replying, in case you do !

Really

Re: Hard/Soft realtime systems

QUESTION ON "Research"

EMSDN.COM

It's a question for 'Hard/Soft realtime systems' under 'Research'

Research

Hard/Soft realtime systems