7 ms·
As 0days get meaner, Google defenses increasingly outpace Microsoft
- EpicEng 12y agoSo... the article makes a direct comparison between updating a centralized server and updating millions of computers individually. And oh yeah; the latter is an OS patch. I'm not defending any specific instance here, but in general, I think it is reasonable to expect a Windows patch to take longer to develop and deploy than an update to Google Apps.
- CJefferson 12y agoYes. Let's consider how well Google do with Android. Even Android on Nexus devices, where they (in principle) should have much more control. Their update speed, and how long versions get updates for, is still far behind Microsft.
- dmix 12y agoHow much of this is due to the UX of updating mobile operating systems vs software? OTA updates on Android phones means downloading (often 250mb) files -> asking the user to reboot to recovery -> installer takes a few minutes -> boots back into the phone. Most people keep their phone on 24-7. So this is not as simple as updating software apps in the background w/o reboot.
- throwaway5752 12y agoA lot of the Windows security updates require restarts. It's not a dissimilar problem. And OS patches have to happen on shared systems, Exchange servers, Sharepoint servers, domain controllers, etc. Uptime is more important on those than a cell phone by any objective measure. edit - Also, Microsoft has complicated relationships with many large vendors and probably tests on a LOT of permutations for each patch vs whatever Google does for Android. I am surprised to find myself sympathizing with Microsoft.... that being said, this is digressing. MSFT should have a universal fast track for 0days and decouple it from the standard upgrade process. 90 days should be enough, and have no real problem with Google's actions.
- dmix 12y agoIndeed, but this must be considered if someone is going to critique Google's speed at updating Android vs desktop operating systems. The updates need to be stable and important if the user is going to be walked through this disruptive process. Additionally, Windows OS has a long history, is much more stable, and Microsoft controls the majority of the codebase unlike Google. Android's codebase is built on top of quite a rickety legacy foundation of old patchworked linux kernels that have to be in sync with drivers from proprietary manufacturers like Qualcomm - although is has improved a lot in recent years. Yet, I do agree Google/Android vendors should be pushing updates at a much higher frequency.
- luuio 12y agoNot disagreeing with what you said, just want to point out that Microsoft needs to deal with even more driver from proprietary manufacturers :-)
- dmix 12y agoGoogle has a critical dependence on a few very slow moving manufacturers before they can do each update. And their vendors aren't nearly as capable or incentived as Intel or Nvidia to release updates quickly - since their arrangement is a fixed-hardware deal with particular phones. Not for supporting x operating system on a variety of hardware, with competitors at their throats.
- luuio 12y agoI cannot think of a company that works with Android and not with Windows. Again, agree that Android has to support a lot of hardware, but Windows needs to support even more.
- snowwrestler 12y agoOTA updates work exactly the same way on iOS devices, but a far higher percentage of iOS devices are running the latest version of their OS, than are Android devices.
- dmix 12y agoWe're talking about patch-based updates that happen on a frequent, not full OS updates. OS updates on Nexus devices always happen immediately just like iPhones, which is what we're talking about... Non-Nexus devices distributed with OS modified by 3rd-party vendors and/or cheap/low-powered Androids are the big reason why Android phones stats show a lack of updates. It is not because of Google - which does push out OS updates in a timely way. Yet Google still releases patch/delta very infrequently which is what we're comparing vs Windows.
- droopybuns 12y agoThis is not apples to apples. As soon as you apply the same ecosystem concepts (end points not owned and operated by the company) google has the exact same problems. Reminder: Silent closure of android bugs over the holidays. How many beers did the google pr team buy Dan Goodin? http://www.reddit.com/r/Android/comments/2qjkuf/someone_silently_closed_37_1952_of_android_bugs/ http://www.reddit.com/r/Android/comments/2qjkuf/someone_sile...
- Grazester 12y agoI don't see how that article made Google look particularly good. Maybe we read different articles. For whatever reason I just can't help but think 3 months was enough time for these bug to get fixed. Didn't Microsoft want Google to delay the announcement until their patch Tuesday when they usually release their updates? If an update is critical(which I guess wasn't) then why not push it before hand if you have it already?
- 100timesthis 12y agopatch Tuesday it's one of the best thing of Microsoft. It's predictable so so it's possible to schedule the patching in advance. The patch isn't something magic that appear in the system, large organization need to be sure the patch isn't interfering with their other processes.
- Grazester 12y agoNo large organisation do auto update so does it matter if say a critical patch is release so their IT department can test that right away and push it?
- droopybuns 12y agoLogistics and supply chain are real fields. Distributing updates that don't break one off oem combinations of hardware required testing? Apply your logic to android. The patches exist in AOSP. Why don't they get to the devices? If you don't own the servers that get the patch, your testing obligations are much higher. The Google team is indifferent to realities that they aren't exposed too. This behavior is childish and rude- and it isn't going to be the kind of behavior that the EU is going to look kindly on when they press their anti trust case.
- jeremy_wiebe 12y agoIn addition to the already noted comparison between a web app patch and OS patch it's interesting that the organization that has found these two security vulnerabilities in Microsoft software is (partially?) funded by Google.
- landr0id 12y agoGoogle's Project Zero has actually found quite a lot of vulnerabilities in commonly-used software. To date they've publicly disclosed 107 vulns, most of which include a PoC: https://code.google.com/p/google-security-research/issues/list?can=1&q=&colspec=ID+Type+Status+Priority+Milestone+Owner+Summary&cells=tiles https://code.google.com/p/google-security-research/issues/li...
- norswap 12y agoThe article makes an interesting point I hadn't considered when I first heard about Google disclosing MS's vulnerabilities. And then I thought how rare this was (and not because I'm a person of great insight) and it made me sad.
- sarciszewski 12y agoWhat exactly is stopping Microsoft from issuing patches as soon as they're ready, and sysadmins configuring their machines to only look for said updates on the second Tuesday of every month? This model works for other operating systems.
- CJefferson 12y agoBecause there are teams of people who jump on patches as soon as they are released, backward engineer what they fixed, and use that to produce attacks. Of course for bugs which already have public 0-day attacks, that isn't a problem, but the majority of bugs are not publicly known at fix time, and it is important that everyone gets the patches applied ASAP after release, hence the well-defined day when patches are released. At my work that day is purposefully blocked off to get patches applied, and be prepared for any fallout.
- sarciszewski 12y ago> Because there are teams of people who jump on patches as soon as they are released, backward engineer what they fixed, and use that to produce attacks. This should serve as an incentive to stay on the bleeding edge of security updates. Make automatic updates automatic without requiring a restart every time. Not an incentive to sit on our hands and wait until the second Tuesday of a month. The term Exploit Wednesday was coined in response to Patch Tuesday. :) Ultimately, I think this change from a monthly security release pattern to a "as soon as it's ready" security release pattern is what needs to happen. For the good of everyone. Then if people get 0wned, it's their own damn fault.
- cwyers 12y agoEveryone pointing out Google's Android update policy of "buy a new device, asshole" isn't wrong. But I think an honorable mention goes to Chrome's update policy of "who gives a shit if something breaks anyway." The whole article is laughable.
- spacemanmatt 12y agoApple stops supporting older iOS devices, too. But I agree, Chrome quality kinda goes up and down.
- serge2k 12y agoWhen the devices are so old and far behind that they can't reasonably run a new OS
- spacemanmatt 12y agoMy take: 90 days is a reasonable amount of time for Microsoft to get a fix out. It's not a reasonable amount of time to get a fix out and get it deployed to a reasonable percent of the user base. The article is not just saying Google can fix their web applications faster. They're saying that is the whole point of making technology choices like webtop versus desktop. Further, after the decades of outright disdain to bother releasing a stable, much less, a secure product from Microsoft, I'm not in a hurry to grant them leeway on this. If they published a counterpoint explaining why they need so much more time, I'd give it a fair reading.
- iolothebard 12y agoThe fact you have an opinion on what's reasonable for MS to do is fucking comical. How much do you know of the inner workings of MS? The great internet arbiter.
- spacemanmatt 12y agoMicrosoft's competitive problems are not my concern. Other platforms do better than 90-day turnaround. No "inner working" knowledge is needed here.
- jlarocco 12y agoNot to defend Microsoft too much, but that's a bogus comparison. And besides that, I don't even know what the point is. It's faster fix and deploy websites than it is to fix and deploy operating system patches? Well no shit.
- AceJohnny2 12y agodeployment != fix However, Microsoft does have to contend with a much larger and more legacy diverse codebase, and astronomically higher risk of regression.