WebGL in Enterprise Web Apps? Uh, no. [Update: make that: not yet]

When a big part of your business is making 3D product representations available to as many people as possible, you pay particular attention to technologies that might help you achieve that goal. WebGL is a promising new technology that can bring 3D rendering into the browser as a native feature. And if you follow the hype, you might think it was already widely deployed.

However, uptake among web consumers does not necessarily predict availability in the enterprise. To figure out whether WebGL is an option for our B2B customers to deliver their sales and marketing messages to enterprise users, we instrumented the login page of our content download portal. Over the course of a month, we tested WebGL availability for 1,872 visitors’ machines.

Most of the people coming to our content portals are field sales and marketing teams, and channel partners, for large high-tech companies. Over the sample period, most of the visitors were from companies that sell computer networking, hardware, and infrastructure products. Of all the enterprise users, you would expect that these are the most likely to have access to cutting-edge technology.

To cut to the chase, the portion of these users with availability of WebGL support in browsers is just 24%. (If you only consider those with hardware-accelerated 3D, the number would be more like 19%; I’ll explain this below.)

We classified users into 3 groups:

  • (Y) Those who have a browser with WebGL available (24%)
  • (G) Those with a browser that supports WebGL, but it is disabled (often because they have a graphics card that cannot support WebGL) (26%)
  • (N) Those with a browser that has no idea what WebGL is (50%)

Our data shows that Chrome users are pretty much all up-to-date, which is very significant because the current version (18) of Chrome includes a software rendering solution for WebGL. That means that people who would not have been able to view WebGL in version 17 of Chrome (“G”s) now can (making them “Y”s).

Since Firefox includes no such software pipeline, we can guess the number of Windows users who have hardware that does not support WebGL, even though their browser does. It turns out that is a whopping 57% (115 Firefox “Y”s, vs 147 firefox “G”s). This tells us three things:

  • Most enterprise users do not have up-to-date 3D accelerators in their Windows computers
  • The addition of a software engine is a critical feature for success in the enterprise
  • Enterprise WebGL apps should be tested and optimized for software rendering, because a majority of users will see that version

Digging into the data a little deeper, we see that Apple has a pretty impressive foothold these days. In our sample, 24% of users were on Mac, and 4% of users were on iOS, leaving just 72% of users on Windows. We do not know whether this reflects more adoption of Apple as enterprise hardware, or is just reflective of field sales forces using the best computer for their needs, regardless of what the IT department prefers they use.

While pretty much every Apple computer and tablet has 3D acceleration hardware built-in, it turns out this doesn’t bolster WebGL’s number much. The problem is that the default browser for both Mac OS X and iOS is Safari. Safari knows what WebGL is, but it has it turned off. So 61% of Mac users are “G”s, which could become “Y”s if they just knew what hidden box to check.

If we look at the breakdown of people who have a WebGL browser, we see that, across operating systems, 54% of those users are on Chrome, and 46% are on Firefox.

If we look at the breakdown of people who cannot use WebGL, we see that, across operating systems, 60% are using IE (which has no WebGL support at all, nor plans to add it [Update: see below]), 25% are using Safari (on Mac and iOS), and 13% are on Firefox. Only 1% are on Chrome, and they’ll move into the “Y” camp as soon as they get up-to-date.

So what has to happen for WebGL to become a viable enterprise option? The quickest path would be IE adding WebGL support, but that seems very unlikely to happen. [Update: see below.]

Firefox has a stronger presence in the enterprise than chrome (which is a little surprising, actually). So getting a software rendering solution into Firefox would be a big help.

Apple can change the default for Safari from “Disabled” to “Enabled” with a tiny software update. That would bring in just about all the Mac users. Enabling WebGL in the version of Safari for iOS would bring in those users, as well.

Together, a software renderer in Firefox plus Apple fully embracing WebGL would bring support to 55% of our enterprise users. That leaves just 45% of users on Internet Explorer, with no WebGL in sight. Is 55% coverage good enough? No, not for Kaon it isn’t.

My conclusion is that while WebGL is very cool, Microsoft’s refusal to support it will kill it as a web technology in the enterprise. [See update, below]

This does not mean WebGL has no place in the enterprise, of course. People have to install special software to run enterprise applications all the time, and by installing Chrome, they could get access to this new technology. But if you are designing web content that you want enterprise users to see in the browser they are using today, then WebGL is not a viable option, and there does not appear to be any way it will become viable in the foreseeable future.

Update: July 30, 2013

After stating categorically that they will never implement WebGL, Microsoft surprised everyone by not only deciding to include WebGL in version 11 of IE, but to back port IE11 to Windows 7, and to provide a software rendering option for people without a trusted hardware pipeline. This is huge. Judging by visits to the kaon.com web site, it looks like only 20% of enterprises are still at Windows XP. That means that more than 80% of Windows users are on a version of the OS that supports IE11. Furthermore, MS has shifted to a forced-update approach for the browser, which means that provided you are not still on XP, then you will almost certainly be running IE11 shortly after it is released.

It is hard to predict timing, since MS is being cagey about exactly when IE11 will move to general release. However, it is looking like sometime in 2014, WebGL may reach critical mass in the enterprise.

About kaonalphageek

CTO and co-Founder of Kaon Interactive Inc.
This entry was posted in Geek Talk and tagged , , , , , , , , , , , , , . Bookmark the permalink.

16 Responses to WebGL in Enterprise Web Apps? Uh, no. [Update: make that: not yet]

  1. Pingback: The SILLY APPS « SILLY APPS

  2. Pingback: WebGL around the net, 17 May 2012 | Learning WebGL

  3. Benoit Jacob says:

    The problem is not with hardware not being capable of supporting WebGL, it’s with people having too old graphics drivers. The vast majority of the ‘G’ category is people who are just a driver update away from getting WebGL in supporting browsers. See http://www.khronos.org/webgl/wiki/BlacklistsAndWhitelists

  4. kaonalphageek says:

    I don’t have any evidence that says whether that is true or not. It might be. I do know that many of our customers are mobile, which means many of them have mobile Intel chipsets that are blacklisted regardless of the drivers they are using.

    But the point is academic, since:
    1. Enterprises, as a rule, never update graphics drivers; and users do not have the privileges (or often the skills) to update the drivers themselves; and
    2. 45% use IE, which makes pretty much everything else irrelevant.

  5. Benoit Jacob says:

    All Intel graphics chipset since the Intel GMA X3000 from year 2006, are capable of fully accelerating WebGL. That’s the overwhelming majority of Intel-based PCs and laptops sold over the past 5 years. Older Intel chips such as the Intel 945 still allow WebGL with some software fallbacks implemented in the Direct3D driver.

    All it takes to support WebGL on Windows, is Direct3D 9 support.

    So the problem is solely your 2 points below — drivers not getting updated, and IE not supporting WebGL. There is no hardware problem here.

  6. MrWebGL says:

    Chromeframe and IEWebGL makes running WebGL on IE as simple as installing any other plugin (unity3d etc). Drivers get updated as people buy new computers. The trend of IE is a steady slope downwards.
    It’s funny that you are the CTO of Kaon Interactive, whose 3d solution builds on Java+applets. It’s understandable if you feel like the rug is being swept away from under your feet with WebGL. 🙂

    • kaonalphageek says:

      We actually have a WebGL solution pretty much ready-to-go. Right now Java gets us onto 95%+ of the desktops, so that is the best solution, but the minute WebGL passes Java in availability for our customers (who are all B2B), we’ll switch.

      I wish I could share your optimism about plugins saving the day, but I’ve been hearing that argument in the 3D world for 15 years now, and yet no 3D plugin has ever come close to the universal adoption. Plugins have never been a good option in the enterprise.

      Chrome frame has been available for quite a long time now, but we’ve seen no evidence of enterprise customers buying in to that.

      IE’s slope is downward, but it is not very steep. If getting to 95% availability means waiting for IE to reach 5%, that doesn’t sound like a very good strategy.

  7. I’ve created a site with a bit a broader sample base (10 million visits, 6 million visitors and counting) and the numbers are up way more than in your stats (http://webglstats.com/)

    The large discrepancy must have something to do with the kind of hardware that enterprise users get to do their job (low-end craptops and crapbooks), the kind of software update philosophy (never) and the kind of browser choice/update (none, never).

    The browser is really just the smallest missing puzzle piece. You simply can’t deploy 3D content to any enterprise customers unless you make them buy machines to do it. I’ve been there, I can confirm it, it’s the same all over Credit Suisse, UBS, Accenture, etc. These are not 3D ready machines.

    • kaonalphageek says:

      Yes, I referenced your site in the post. The numbers you have are from a user type very different than the people we are reaching with our B2B sales and marketing pieces. (That’s pretty clear from the list of sites you reference which have the tracker installed.)

      We’ve been supporting these users for several years now. Our enterprise desktop platform includes a software rendering engine (the same one we use in web applets), and we design the content so users who do not have access to Hardware GL get the software 3D in a resizable window. (3D performance can be just fine in software, if you keep the pixel count down.) Also, keep in mind that the same machines that don’t have 3D hardware also have 1024×768 screens, so the pixel count never gets that high anyway.

      We’re really impressed with the software engine in the latest version of chrome, and I think that it might be a great option for enterprise users who are allowed to use chrome. Unfortunately, that doesn’t hit a relevant market share for outbound marketing. But, as I noted at the end of the article, we think it might make sense for an internal sales force, if you can get IT buy-in to deploy chrome as a platform for that purpose. (Or, at least, not stand in the way of users installing it themselves.)

      • Well, the other point is browser adoption. It’s fairly certain that iOS and Android native browsers will acquire WebGL in the near future (before the year is out or early next year I’d wager). Win8 phones wouldn’t support it (and they also exclude other browsers), fortunately they don’t look to be grandly popular anytime soon, if ever.

        I think it can be expected that Safari will flip it on by default once Apple has implemented GL_ARB_robustness on their desktop drivers, and that’ll likely happen very soon.

        Firefox could contribute a goodly chunk to the market share if they’re less conservative in regards to the device/driver blacklist (chrome is much less conservative). However for that to happen Firefox will have to put tabs into their own processes, I believe that is already being worked on, so I’d expect Firefox to draw even with WebGL support with Chrome fairly soon.

        IE is a big problem of course. But if you look at this problem it actually has two aspects. The first is if/when Microsoft will add WebGL, the second is how long it would take to proliferate trough their IE userbase. Fortunately (albeit not very satisfactory) the answer to the latter question is fairly easy to formulate. It would take them years (and in the case of enterprise customers probably decades) to deploy a new Browser version. So even if Microsoft implemented and released WebGL right now, you might not see significant browser support on Enterprise IEs for in the next one or two decades.
        On the other hand, if Microsoft wants to stay relevant in the browser wars (which they’re loosing right now), they will have to master version conversion better and be more aggressive about updates. And the more market share they loose, the more aggressive will they pursue strategies to catch up to other browsers, which will include faster updates and perhaps WebGL. The question if/when Microsoft introduces WebGL and manages a faster update cycle is therefore rather just a matter of when they’ve reached a certain pain threshold and if they will have the capacity to react correctly to stay relevant.

        On the other hand, IE is hemorrhaging users year after year, and this isn’t quite a linear function. It is entirely possible that with a sustained failure to draw even with other browsers, IE could fall off a cliff where users convert to other browsers en-masse.

      • kaonalphageek says:

        Great points, all. The other part of the mobile puzzle is performance. We’re seeing really awful performance in WebGL in Android right now (compared to our native GL-ES-based iOS apps on iPad/iPod/iPhone). I’m sure that this is just a matter of time, to wait for the hardware to speed up.

        WebGL in Safari on iOS would be fantastic for us (right now, our iOS 3D solution requires a native app; we’d love to be able to provide in-the-browser 3D for iOS, like we do on Desktops using Java applets). I read an interesting blog post by someone who figured out how to turn WebGL on in the WebUIView, using one (private) API call, so it’s clearly an issue of Apple just being conservative. Given that there are innumerable web pages that kill Safari iOS by using too much memory, I’d guess that the conservatism is about exploits, not stability. (They have WebGL turned on in iAD’s, but of course, they vet those, so exploits are not a concern there.)

        On the desktop, it’s possible (unlikely, but possible), that enterprises buy into the idea of Metro’s IE which has a completely crippled experience in exchange for very high security. (No ActiveX, no silverlight, no flash, no java, no fun.) In that case, there will be no way to deliver 3D using any technology, which would be a bummer, but it would re-level the playing field for WebGL.

        Or, perhaps, Microsoft will come out with a completely non-standard way to do 3D (vis. VML vs SVG) and we’ll all just have to do WebGL on everything else, and Microsoft’s proprietary thing on IE. That would be unfortunate, but very consistent with Microsoft’s past business practices.

  8. Antonio Salvemini says:

    Hi Kaon people,
    I am Antonio Salvemini –
    co-author of Cycore Cult3D, co-author of mental images’ RealityServer, worked at mental images and NVIDIA Marketing (in fact, I know you from my past competitive analysis ).
    When it comes to 3D technologies, I have seen and done it all. I’ve been pioneering 3D on the Internet since mid-90s.
    I am currently keeping an eye on WebGL and I will probably select it for my next projects, but my conclusion is exactly the opposite as yours. Since WebGL’s adoption is relatively, especially on mobile devices (despite various announcements and brave early adopters such as Sony), applications such as sales support and promotional marketing are currently a no-go. But that’s what I would largely define as B2C marketing or at least mass-marketing, where the ability to reach out to as many users as possible is mission-critical.
    However, one thing is to deliver a mass targeting product, and quite another is to deliver a very specific, problem-solving solution to an enterprise that can fund your work and will easily fill all the gaps to make it happen, whether that means an HW upgrade or installing new software.
    And by the way, outdated HW is a good thing: it means there’s market for HW vendors and you can use them as multiplicators for your propositions.
    And then again, in my experience, a lot of things can be done by creating bottom-up demand, and the technologies that we marketers have today make this easier than ever (viral distribution, building consent, publicly commenting to large companies etc…). So we’ll see, but I believe that WebGL has the potential for not turning into a fad – it will be up to developers and content providers to make it fly.

    PS for you stats are the result of your own positioning and targeting, minus everything falling off your market share. For “unfiltered” stats I would suggest http://www.w3counter.com/globalstats.php

    • kaonalphageek says:

      In my experience, getting an enterprise IT department that dictates the use of IE to switch to a different browser is pretty much impossible. Even getting some corporate IT departments to *allow* the use of a different browser can be a challenge. But, hey, I hope you’re right! Our WebGL implementation is all done, and we’re really happy with it on the small subset of browsers/hardware where it works! (Except that we cannot get pinch-zoom to work, since the browser eats that gesture, which leads to a real problem with the user experience.)

  9. kaonalphageek says:

    If you are lucky enough to be using a WebGL-capable browser, have a look at this demo we just finished:


    • Antonio Salvemini says:

      It works very well and out of the box, using Google Chrome and an NVIDIA Quadro Fermi card. All animations are extremely smooth in full screen. Cheers

Tell us your thoughts

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s