MS Responds To Vista's Network / Audio Problems

quirdan writes "With the discovery last week of the connection between Vista's poor networking performance and audio activities, word quickly spread around the Net. No doubt this got Microsoft's attention, and they have responded to the issue. Microsoft states that 'some of what we are seeing is expected behavior, and some of it is not'; and that they are working on technical documentation, as well as applying a slight sugar coating to the symptoms. Apparently they believe an almost 90% drop in networking performance is 'slight,' only affects reception of data, and that this performance trade-off is necessary to simply play an MP3."

Comments

Back in 1994...

[Anonymous Coward]

Back in 1994, I bought a Power Macintosh 7100. One of the first PPC chips, about 66MHz, and running a positively archaic operating system.

I still have the machine, and drag it out from time to time. When this story broke, I pulled it out of storage to test it, and see how it compared. With a 10/100 ethernet card in, running the mac's System 7.5.3, it could successfully play an MP3 while transferring, and it made no difference whatsoever to send or receive speed over the network.

Take note Microsoft: 1994, 66MHz, System 7.5.3, more than 13 fricken years ago.

"It's not a bug, it's a limitation."

[John Jorsett]

That was the response of a MS tech regarding a defect that a bunch of us found in one of their C libraries some years ago. They must have had that guy train his successors.

Re:"It's not a bug, it's a limitation."

[iluvcapra]

There is this sort of undercurrent in a lot of Microsoft literature on MSDN, Channel9 and other sources (see this in particular); many units in MS seem to take it for granted that "computing" is essentially an activity of programmers, and that end users need not be bothered with it. Sure, end users use computers, but really all they do with them is stuff they could've done without them, just faster (according to MS).

Since an operating system is a "computing" product par excellence and really has no relation to a practical end-user process (by their reckoning), Microsoft only indifferently supports its operating system for end users, and primarily targets its attention on getting developers to make the switch. They believe, for good reason, that if they get the devs to build on Vista, then the end-users will just follow the applications, and that they won't really need to market the OS. Or, for that matter, even spend too much energy supporting it, since performance and reliability are always secondary to compatibility, which the developers lock the end users in to.

From the horse's mouth

[stinerman]

FTA:

"The connection between media playback and networking is not immediately obvious. But as you know, the drivers involved in both activities run at extremely high priority. As a result, the network driver can cause media playback to degrade. This shows up to the user as things like popping and crackling during audio playback. Users generally hate this, hence the trade off."

Granted, I don't want my audio stuttering, but the idea that the CPU can't keep up because of file transfer is insane. Maxing out an ethernet connection doesn't take much CPU. Even if we put the audio at a very high priority, I don't see how that would immediately degrade ethernet performance by 90%. I could accept no more than about 5% in a worse case scenario.

To be fair if I renice rhythmbox to 18 and transfer a file, things go to hell. Renicing to 10 clears it up. I saw no degradation of speed. Apparently Debian can do file transfers at full speed while playing an mp3 on a rather old PC*. Something isn't right here...

*Athlon XP 2400+, 1GB DDR

Re:From the horse's mouth

[Xtravar]

Generally, audio goes through many buffers before being sent out to the device. In Vista, perhaps all audio is streamlined as high-priority.

For example, when audio recording, you don't want to use Microsoft's typical sound system - you want to record using ASIO which goes through less buffering and latency. If you record using the regular sound system, you end up with perhaps 100s of ms of lag, which is a bitch when you're trying to record to a metronome.

As some AC above noted, Linux only has a direct audio IO path when using jackd. Otherwise, everything is buffered a plenty.

So I think it has nothing to do with CPU power, and more to do with "Vista is a real-time multimedia machine!" When you're interrupting a LOT to be attentive to the audio device, this is going to interfere with the network, whereas if you just interrupt less regularly but send larger amounts of buffered data you don't have that problem.

*Fair warning, my facts may not be 100% accurate, but I think this is the gist of the problem.

FTA

[flummoxd]

"In certain circumstances Windows Vista will trade off network performance in order to improve multimedia playback. This is by design."

I know we've been over this before. But for whom are we 'improv[ing] multimedia playback'? Is it really an issue in 2007, to perform a network transfer and play an MP3? Or is it Vista's "secure audio path" that is responsible for this? Remember, this is the same Vista that polls your hardware every few ms to check if you're playing 'premium content'.

I know not everything bad Microsoft does is done with forethought and malice (..) but really now. After reading the 'cost analysis of Vista content protection', can you not understand the apprehension? If some "multimedia" (albeit not 'premium content', but who's counting) is played, other parts of the system deliberately go into a 'limited' state? After reading that, does it sound like a bug to you?

"But as you know, the drivers involved in both activities run at extremely high priority. As a result, the network driver can cause media playback to degrade. This shows up to the user as things like popping and crackling during audio playback."

I call shenanigans.

Even if this is a legitimate "bug", i.e. the Vista testers were actually experiencing crackling audio while performing high bandwidth network transfers, who made the conscious decision to throttle the *network* instead of fixing the audio path and audio drivers? Windows XP had no problems performing high-bandwidth transfers and using the sound simultaneously. Besides normal operating system scheduling there was no 'throttling' of any device A when any device B activates. This is Vista content protection backfiring, plain and simple.

Re: Deployment

[TaoPhoenix]

Maybe they don't care at all about deployment of Vista.

We harp on MS a lot, but they ARE clever in certain ways. Suppose someone is thinking Big Picture in some kind of twisted sense. They can play a variant of GoodGuy/BadGuy by having a "Sacrificial OS" every 8 years. They're somehow getting us to pay for their beta testing. They HAD to get Vista out, period, and rely on their patented brand of bluster to get through it. They were getting serious heat from inactivity. I bet someone got utterly crushed when they had to switch codebases during that dev setback.

I barely heard of Win Me - consecutive tips told me to get Win2000, which lasted me through 2.5 OS changes from MS. Then in the early days, I saw a lovely crash&burn act on XP *SP2* until everyone repaired their firmware. I even had some flash devices that I had to return until the factory shipped ones with newer firmware.

Now XP is their heavy duty workhorse while they experiment with their new codebase. Suppose just for a moment that Vista NEVER works... but what they learned from Vista SP1 gets applied to Windows 7 (anyone got a codename yet?). Then maybe by 2010 all the results of history on the media scene will be in, maybe they will back off from DRM, and take some other focus. If they don't screw it up, Vista will be that smile in techie's forums, Windows 7 will be the new 8 year workhorse, and off we go ever after.

Having cash flow the size of a country must be fun.

All MP3 Players?

[Nom du Keyboard]

So does this affect all Windows media players (e.g. WinAmp), or just WMP? Could be a great argument to jump ship to non-MS software.

Re:REally?

[gad_zuki!]

Maybe we should start using the slashtards tag. Did you even bother to click on the link? Hell, the article is written in the ADD-style of "dummy quotes." The author doesnt even present the full email! In short, what was left out is that MS has acknowledged the bug but the tech wrote that there is going to be some kind of performance hit. Its not like MS wrote "THIS IS NOT A BUG. GO AWAY." Contrast:

"Please note that some of what we are seeing is expected behavior, and some of it is not. In certain circumstances Windows Vista will trade off network performance in order to improve multimedia playback. This is by design."

"In most cases the user does not notice the impact of this as the decrease in network performance is slight. Of course some users, especially ones on Gigabit based networks, are seeing a much greater decrease than is expected and that is clearly a problem that we need to address."

In other words they see a bug especially on gigabit connections.

Now back to yoru regularly scheduled bitching and "ZOMG my calculator gets better performance" fact-free discussion.

Re:missing tag?

[Yvan256]

Microsoft should just tell the RIAA/MPAA to go fuck off. Seriously.

I just took a different route and told the RIAA/MPAA to go fuck off by buying a Mac mini.

Say what you want about Apple but at least they're not bending over every time the RIAA/MPAA asks them to do something.

Had to uninstall Vista from the 2 newest laptops

[GoldTeamRules]

We bought at our company. It was just ridiculous the number of times programs crashed (photoshop cs3), how slowly development environments ran (Java, Eclipse), and how terrible disk I/O was.

Remember, this was supposed to be an UPGRADE. Honestly, it is just terrible. Vista on a laptop is simply awful. These were brand new HP laptops with 2GB of RAM.

Vista offers nothing. It is an utter waste of time to attempt an upgrade at this time. With Vista and IE7, the shine is definitely off of MS. There is nothing in the MS product roadmap that is even remotely interesting to me at this point.

MS competitors have never had a better time to take advantage of MS market position than they do now. The hole is wide open.

Re:M$ expected behaviour!

[Richard Steiner]

I play MP3 files in the background all the time on my OS/2 box (a Micron PPRO/200 tower w/192MB built in late 1996), and it has no measurable impact on network activities. I usually use Z! as the player.

Re:REally?

[Kjella]

Well, from what I've understood a TPM chip isn't required to run Vista (though it's a requirement for some of the "Vista-ready" stickers I believe). With no TPM chip this can't kick in, so can anyone with a mobo without TPM test it?

Re:missing tag?

[TDRighteo]

Yes, it still happens.

The original reports noted that foobar2000 was just as affected as WMP. The problem occurs when the audio driver is in use. Interestingly, pausing foobar2000 seemed to release the audio driver (network performance went back to normal) while pausing WMP did not. VLC performs in a similar manner to foobar2000, although bypassing the audio device (decoding straight to a null device) results in no slow down.

So, no, it's not the checking for DRM while unwrapping the MP3 like you suggest. You can do that quite happily via VLC, provided you don't intend on HEARING it. ;-)

Re:New OS has old problems

[coryking]

Actually, the TCP/IP stack is a rewrite. Assuming this bug is somewhere in the TCP/IP stack, this is a prime example of why you should *not* rewrite.

Re: Deployment

[gatesvp]

Good "conspiracy theory". Ever heard of Singularity? Whole OS written in C, Assembler and Managed .NET. They've end-of-lifed FoxPro and VB6, I'm sure that ASP will dying. They've started moving big chunks of Office 2007 to .NET so it's probably just a matter of a few years before they're ready to dump everything into managed code and start rolling out Singularity (Windows 2010?).

You're really not that far off, people have been "waiting" for Vista, but this is really a throwaway OS, nobody is using it and it's not like business is "clamoring" for even this version. Heck many Enterprises have just finish rolling out XP. The new WPF and WCF will surely be functional under Singularity, and Enterprises are just now moving to Managed Code applications (check out the market for ".NET developers"). MS won't die away if this Vista "fails", so we're probably all looking at a Managed Code future in 2010 or 2011 :)