[app_Developer]

Don't know why you just focus on the RT scheduler when it is the RT OS that we are talking about.

Because that's what the RT in RTOS means. It's the scheduler. Everything else is other aspects of the kernel. The aspect of the kernel that I think is inappropriate for this application is the RT part!

BlackBerry made the decision for a few reasons. Take the time to read why BlackBerry choose QNX for security.

https://www.troopers.de/media/filer_...l-aint-one.pdf

https://labs.mwrinfosecurity.com/***...2016-03-14.pdf

Also because of the design, there are about 1/4 of the kernel calls needed to function the same with QNX which means less possible vulnerabilities.

Like I said, I'm not disagreeing that security is an important strength of the QNX work. It's unfortunate they couldn't marry that expertise with deeper experience in smartphones. They could have written a new scheduler. Linux is been used with a few different ones now, including at least one good RT scheduler.

[app_Developer]

LOL. So in your opinion, BlackBerry should have gone linux like the successful Open webOS, Jolla Sailfish, Samsung Tizen, Amazon Fire OS, Ubuntu Phone and Firefox OS.

Yes, because I think for BB10 to have any chance, it needed to be earlier and it needed to have more immediate access to the latest SoCs (again I mean early on). Linux would have helped with both IMO.

They could still have build BB10 on top. Qt already ran on Linux. All SoC from Qualcomm support it out of the box. You can still lock down the boot loader. If early enough, they could even have omitted the whole android runtime altogether.

[DrBoomBotz]

But could BlackBerry release QNX without it being a RTOS? No. QNX is what it is.

These two properties in question can be discussed together or separately.
Without intimate knowledge of QNX internals I would guess that BlackBerry(when they had the engineering resources) could have made QNX non RTOS.
It could also be made into a monolithic kernel.
The former would likely be a kludge and the latter is just crazy.

[Bluenoser63]

These two properties in question can be discussed together or separately.
Without intimate knowledge of QNX internals I would guess that BlackBerry(when they had the engineering resources) could have made QNX non RTOS.
It could also be made into a monolithic kernel.
The former would likely be a kludge and the latter is just crazy.

And they could still be working it. They had QNX as a whole and added a GUI and other processes on top of it. That would be foolish.

[Bluenoser63]

Yes, because I think for BB10 to have any chance, it needed to be earlier and it needed to have more immediate access to the latest SoCs (again I mean early on). Linux would have helped with both IMO.

They could still have build BB10 on top. Qt already ran on Linux. All SoC from Qualcomm support it out of the box. You can still lock down the boot loader. If early enough, they could even have omitted the whole android runtime altogether.

Yeah. Like Linux helped all other other Linux phone OS versions? And it wasn't the SoC or QNX or RTOS that caused BlackBerry problems. And most of those other Linux systems were still be worked on by the time that BB10 was released. So having it earlier wasn't going to happen with your wishes.

[Yasch22]

Regarding QNX revenue, John Chen said on BNN's The Disruptors last June that FY/15 revenue was "over $100m." BNN - Watch TV Online | The Disruptors for Thursday, June 18, 2015
My guess for FY/16 is that it reached $130m, as Chen said core software revenue grew 24%, and QNX probably performed slightly better than BES-Good.

In FY/15, QNX was worth 40% of BB's total software revenue of $250m.

A problem with QNX's prominence in so many verticals, including automotive, is that a lot of companies licensed QNX and developed products for it prior to BB's takeover. Companies like Harman have been developing infotainment systems using QNX for a long time. Also, there was a window of several months in 2010 where QNX was made freely available to all developers, a window that closed the minute that Balsillie and Lazaridis closed the deal. A couple of years later, QNX put out a significantly revised update, and they've been using that as the basis of most new or updated licensing agreements.

The QNX hypervisor is also alive and well.

[Yasch22]

For FY 2015 (March/14 to Feb/15), QNX made "over $100 million" on the "automotive side."
See the first minute (introduction) and/or the final segment of BNN - Watch TV Online | The Disruptors for Thursday, June 18, 2015

[Dunt Dunt Dunt]

Regarding QNX revenue, John Chen said on BNN's The Disruptors last June that FY/15 revenue was "over $100m." BNN - Watch TV Online | The Disruptors for Thursday, June 18, 2015
My guess for FY/16 is that it reached $130m, as Chen said core software revenue grew 24%, and QNX probably performed slightly better than BES-Good.

In FY/15, QNX was worth 40% of BB's total software revenue of $250m.

A problem with QNX's prominence in so many verticals, including automotive, is that a lot of companies licensed QNX and developed products for it prior to BB's takeover. Companies like Harman have been developing infotainment systems using QNX for a long time. Also, there was a window of several months in 2010 where QNX was made freely available to all developers, a window that closed the minute that Balsillie and Lazaridis closed the deal. A couple of years later, QNX put out a significantly revised update, and they've been using that as the basis of most new or updated licensing agreements.

The QNX hypervisor is also alive and well. QNX Auto Blog: Toyota Entune It's the basis of an infotainment system's ability to run apps like CarPlay and Android for Auto over top of the basic QNX unit. Harman put out a YouTube video saying this was how their system operated, and Harman is still most likely using QNX, as they have in the past.

Might have been part of the sale that Harman was free to use QNX code.... even the newer stuff. All BlackBerry wanted back then was a smartphone OS, that no other company could use. What Harman was doing, and the rest of QNX's automotive and other verticals they were involved with.... was peanuts for BlackBerry, they were pulling in $5.5 Billion a quarter in 2010. But by BB10's launch the were burning money and any segment that showed potential was being looked at. Thus Thor dusted QNX off and show us (and more importantly investors) Momentics and started talking IoTs.

Revenues.... hard to say the way BlackBerry keeps changing how they structure Services and Software.

[DrBoomBotz]

And they could still be working it. They had QNX as a whole and added a GUI and other processes on top of it. That would be foolish.

It sure would.

[Jack Chin]

Check the quarterly reports for them and look at the mobile phone division. Samsung is the only one making a little profit.

Apple accounted for 91% of smartphone profits last year

Most Android OEMs are making money on their phones, even the Chinese cheapos. I assure you, they would not lose money on each additional phone manufactured, in perpetuity. As to economic profit (i.e., whether their continued investment in smartphone production is more profitable than anything else), yeah, that leaves a LOT to be desired for the middle of the market, e.g., HTC, LG, and Sony.

(Quote = Bluenoser63)
Take a look at the chart for mobile profit market share and tell me that BlackBerry is going to do better with Android.(/quote)

I won't, because they (almost certainly) won't. But that's a separate issue from your thesis, that BB10 is preferable to Android for the company's phone production. In fact, BB10 was a proven market failure from Day 1. With Android, they at least have a thin chance. Though as with nuclear war, sometimes the only winning move is not to play.

(Bluenoser sez. . .)
And they had to provide at the time an end to end solution which could only be accomplished by making a phone. The only space that they are still living is the enterprise space. They are and always have been a enterprise services company who's phones were the fad to the public.(/quote)

Oh, the irony!

Posted via CB10

[cribble2k]

That post doesn't have any proof that a RT OS (QNX) isn't a good choice for a mobile phone.

Poor market share proves RTOS / BB10 isn't a good choice for a phone.

Posted via the CrackBerry App for Android

[Jerry A]

And they could still be working it. They had QNX as a whole and added a GUI and other processes on top of it. That would be foolish.

Except that's exactly what most people do with QNX (and what it was designed to do).

Granted most of those GUIs aren't to the level of today's smartphones. Busy definitely not foolish.

[PygmySurfer]

Because that's what the RT in RTOS means. It's the scheduler. Everything else is other aspects of the kernel. The aspect of the kernel that I think is inappropriate for this application is the RT part!

The scheduler is just one part of what makes an RTOS "real-time". Essentially, an RTOS is able to process data as soon as it is received, without having to buffer it.

Not sure why people seem to be suggesting BlackBerry should've made BB10 non-RT - RT does not have a negative effect on the performance of phones. Granted, it doesn't really have a huge benefit, either, but it's definitely not detrimental.

[DonHB]

I think that's exactly right. A modern smartphone is very different from the typical QNX applications. I think QNX is excellent for other applications, but a poor fit for phones.

I am still waiting for a compelling argument for these conjectures. What you said regarding MS abandoning the use of a RT OS could be that they were already interested in reusing Windows for their mobile platform.

If BB10 had been built on Linux, I think they would have launched much earlier, they would have reduced the cost of adopting the latest hardware, and they could have had more compelling low/mid end devices. I don't know if all of that would have been enough to save BB10, but I think building on QNX did put them at a serious disadvantage.

I think they should have focused on the integrating the Android runtime with Flow to preserve the user experience. They would have launched much earlier and building on QNX technology would be a long-term advantage.

[DonHB]

What is the impact of an RT scheduler on security?

Neutrino's architecture (no less it's implementation) provides it's security. The potential of Neutrino is in its architecture more than it's RT abilities which range from soft to hard.

[DonHB]

The scheduler is just one part of what makes an RTOS "real-time". Essentially, an RTOS is able to process data as soon as it is received, without having to buffer it.

No

Not sure why people seem to be suggesting BlackBerry should've made BB10 non-RT - RT does not have a negative effect on the performance of phones. Granted, it doesn't really have a huge benefit, either, but it's definitely not detrimental.

True

[Gary Chan HK]

Fortune sat down with Patrick Brady(PB), director of engineering for Android Auto:

How easy will it be for QNX, which licenses its Unix-like operating systems for infotainment and safety functions, to operate on Android? Will QNX be a competitor or a partner?

PB: In many ways, you could say Android and QNX are competitors, but in other ways they�re very complementary.

PB: QNX was built around operating real-time safety critical operations, and Android was built around great connected services and things like that. They�re actually a natural pairing.

PB: If you want an advanced ADAS system, I would not advise you to choose Android today. That�s not what Linux and Android were built for. QNX runs in so many safety critical systems from airplanes to automotive, I think they could be quite complementary, and I think you will see instances where cars run both.

Even though Google�s self driving car project is separate from Android, I wonder if there are any conversations about what the operating system for those self-driving cars might be? Is there talk of bringing the two together eventually?

PB: There is certainly no conversations about merging them or anything like that. We definitely do talk with them. I mean even in self-driving cars, you want air conditioning and things like. So they will need to figure out, or their partners will, a platform to power all of that.

PB: We talk to them, but not in anyway about merging.

[Bluenoser63]

Poor market share proves RTOS / BB10 isn't a good choice for a phone.

Posted via the CrackBerry App for Android

So Linux in the terms of Jolla and the others are also failures because of the Linux OS and Windows 10 is a failure too? The OS isn't a factor in a success or failures of a phone.

[Bluenoser63]

Except that's exactly what most people do with QNX (and what it was designed to do).

Granted most of those GUIs aren't to the level of today's smartphones. Busy definitely not foolish.

I was talking about BlackBerry removing the RTOS and redesigning the OS before adding the GUI part. That would have been foolish.

[Bluenoser63]

The scheduler is just one part of what makes an RTOS "real-time". Essentially, an RTOS is able to process data as soon as it is received, without having to buffer it.

Not sure why people seem to be suggesting BlackBerry should've made BB10 non-RT - RT does not have a negative effect on the performance of phones. Granted, it doesn't really have a huge benefit, either, but it's definitely not detrimental.

Finally someone who gets it.

[Jerry A]

The scheduler is just one part of what makes an RTOS "real-time". Essentially, an RTOS is able to process data as soon as it is received, without having to buffer it.

Not sure why people seem to be suggesting BlackBerry should've made BB10 non-RT - RT does not have a negative effect on the performance of phones. Granted, it doesn't really have a huge benefit, either, but it's definitely not detrimental.

That's assuming it has available resources and isn't blocking because it's already guaranteed execution to prior threads.

But I'm being pedantic. Like most kernel architecture discussions it's academic at best or specific usecase compliant at worst (ie there are corner cases where one architecture choice wrings out better results than another).

20+ years of real-world (tm) usage for general cases hasn't shown any one way to stomp all over another.

I think the reason scheduler discussions are so popular is due to their impact on slow I/O systems (ie platters) with expensive graphic operations (ie desktops/laptops).

[Jerry A]

I was talking about BlackBerry removing the RTOS and redesigning the OS before adding the GUI part. That would have been foolish.

Got it. Thanks for the clarification.

[Invictus0]

Poor market share proves RTOS / BB10 isn't a good choice for a phone.

Posted via the CrackBerry App for Android

Symbian is RTOS and it was quite successful in its time.

In the grand scheme of things though I doubt the kernel matters much, at least as far as sales and marketshare go.

[JeepBB]

Totally agree with you. Just don't recall Android at launch (2011) on the PB. Pretty sure that came later.

It did come later.

IIRC, there was a rush to get the OS2.0 Beta because it, for the first time, contained the Android RT. I can't recall what version of Android was in the ART now, but it wasn't the "current" (at the time) Android.

And here was me thinking that I would never again read another micro-kernel thread ...

[keithhackneysmullet]

The failure of Playbook OS and BB10 aren't in the nuts and bolts of the operating system but in walking away from a already mature ecosystem BB7 and trying to start a new ecosystem from scratch which was suicide. RIM realized this much too late which why they added the android runtime to try to piggy back Google's ecosystem . Google started to require android apps deeper dependency on google services which essentially killed BlackBerrys strategy. In 2011 when the playbook came out Blackberrys brand was starting to become toxic . In 2013 when BB10 launched Blackberrys brand was Ebola level toxic. BB10 was dead on arrival