On the 7th July, Google announced that they are developing an operating system called ‘Google Chrome OS.’
“Google Chrome OS is an open source, lightweight operating system that will initially be targeted at netbooks. Later this year we will open-source its code, and netbooks running Google Chrome OS will be available for consumers in the second half of 2010.”
“…redesigning the underlying security architecture of the OS so that users don’t have to deal with viruses, malware and security updates”
“..Google Chrome running within a new windowing system”
“…Google Chrome OS will run on both x86 as well as ARM chips.”
The world of Internet journalism went mad and Techmeme lit up as a result of the announcement. Thousands of blogs responded too. It was quite an interesting response for what is essentially another Linux distro. Perhaps it reflects the desire for a real consumer-level alternative out there. It certainly proves the power of the Google brand and that could be the most important aspect of the whole product.
I’ve been analysing the official announcement, the blogs and some of the podcasts that are available and made some notes. I’ve pulled the main points out of the announcement above and the one that I want to start with is the one that mentions a ‘new windowing system.’
The standard Linux windowing system is based on a software component called X11. It’s been around since the mid-80’s and it’s responsible for placing graphics on the screen, controlling the mouse and providing a framework for a multi-user graphical user interface. It’s used by almost every Linux distribution and almost every single window-based application. X11 is a heavyweight, memory-hungry layer and it certainly needs sliming down for a netbook-type experience but if the X11 component is taken away as the Google announcement indicates and the Chrome OS succeeds in gaining market share, this could be the worst thing to ever happen to the Linux ecosystem. It threatens one of the biggest opportunities for companies like Canonical, Xandros and the hundreds of groups of people working on windowing systems and apps. It also competes directly with the operating system that Intel is working on Moblin. Traditional Linux distributors have already signed up for that in the hope that it will be picked up by most netbook manufacturers.
It’s not 100% certain that Google is talking about removing the traditional X-Windows layer when they talk about a new windowing system but it certainly looks that way. The good news is that the Chrome OS is probably not intended as a replacement operating system. At least not initially. The idea is probably that it’s an OS option for some early adopters but for most, it will co-exist with a traditional Linux or Microsoft windowing system with Chrome OS starting up first (possibly on BIOS or as a fast-boot Linux ) and traditional Linux and Windows being available in the background once fully booted. Virtualisation technologies could take this further to make multi-OS devices a possibility. Imagine your XP or WIndows 7 operating system sleeping while the Chrome OS remains active for immediate use. Google knows that the world isn’t ready for a switch to cloud computing yet and doesn’t want to upset too many companies with the Chrome OS so a multi-OS netbook seems highly likely as a way to ease the idea of cloud-computing into the market. Look at it as a 10-year plan that will mature with mobile technology and mobile internet connectivity.
Taking away the X11 Windowing system means that we aren’t going to see many local, native applications on the system which makes me think that the Chrome OS is going to be the showcase that pulls the Google application suite together in an HTML5 environment. The start menu will simply be a syncronised bookmark system and cookies and Google Gears will be the method to control sessions and offline activity. Google certainly has a rich set of applications and in recent months we’ve seen many of these applications moving to release status but there’s a lot more that needs to be done. High-resolution location awareness. Stylish applications. 3D support and local network awareness. Will Chrome be able to see shared network drives? Hardware support will be critical too. Linux has never been an operating system that you can rely on to be aware of the most recent hardware and this could be that hardest job for Google. Wifi, 3G, Webcams, Audio, Touch, power control, USB and many other elements of hardware that often uses closed-source code. Its the reason that Nokia and Intel got together and Google will have to reach out to similar companies in order to stay up to date, especially when they are targeting both ARM and X86-based hardware architectures. Maybe Google uses Linux Foundation’s Moblin core and re-builds the UI layer on top. That could be the easiest way to attack the netbook market.
Another element that needs to be introduced before Chrome OS is possible is online storage. Google doesn’t have an online storage product yet although there have been many rumors of a GDrive. Expect this to happen before Chrome OS is launched.
What Google does have available is the idea of single sign-on and 100% provider-controlled security. Automatic updates have helped reduce the problem of security patching and application updates but network-side applications take this to another level. The user doesn’t have to be connected to the internet for updates to download and service packs would be a thing of the past. Applications can even be instantly shut down in the case of an important security issue. With all applications using the same set of cookies, you only need to control one application password. If other internet application providers link into Google’s OpenID system, changing a password across 50 applications would be a one-shot move.
The idea of a third party security concept is going to worry a lot of people. You’ll hear a lot of talk about ‘Big Brother’ but when out into perspective, when analysed for real risk, handing over your security to a third party is nearly always a better option unless you make it part of your daily life to run a security policy. Do you have one? I guess not. I was an internet security manager for over 5 years and I learned very quickly that my personal security was atrocious. I handed my data over to Google a long time ago. A risk remains but it’s much lower than before.
Podcasts
I’ve listened to a number of podcasts that cover the topic of the Google Chrome OS. Here’s a rundown of some you might also want to listen to.
Meet:Mobility. Sascha, JKK and I talk about how Intel have been involved and how it could sit alongside Moblin or even use parts of Moblin (in a similar way that Nokia is sharing device drivers with Intel.) We didn’t have the important information that there’s no X-Windows layer though.
Gdgt think browser only. No native apps. They highlight that there’s no X-Windows layer (on which nearly all Linux apps run.)
Buzz out loud. Why not keep the Android brand? No native apps?
Twit. Ryan Block highlights (again) that there’s no X-Window layer which means most other Linux apps wont work. Ryan also highlights the hardware control issues. Audio, screen etc.
Engadget includes special guest Michael Gartenberg on their podcast who thinks it will hurt the traditional Linux community the hardest where netbooks are a real opportunity. A lively and deep discussion.
Mobile Tech Roundup. Kevin Tofel highlights the difference between web apps and native apps and that web-based apps could improve with HTLM5.
Other points to consider.
- 2010 could be more than a year away.
- Open Source means that there will be another glut of distros based around it.
- App store? With no X-Window layer, the idea of selling applications disappears. It’s all about selling services and content.
- Will it be faster than desktop version? Probably not noticeably faster than a clean build of Windows XP Home.
- Is it a mobile-focused OS? ‘Mobile’ does not equall ‘Cloud.’ Many parts of the world, including the U.S., aren’t ready for a consumer focused mobile operating system. The mobile connectivity costs are too high.
- Touch capability for tablet functionality. This is a UI and driver issue.
- 3D support. 3D is important for modern software aesthetics.
- Techrunch brings up an interesting point. Does bundling a browser into an operating system bring up anti-competitive issues? Probably not when the OS is open and free.
- HTML 5. Will Google re-architect its applications for HTML5?
- Location-awareness. Google has back-end that can determine the location from an IP adress but the browser will need to support the Geo-aware API that Firefox 3.5 uses in order to provide control and capability for other web-based applications. It’s arguable that GPS support would be needed for fine-grain accuracy too.
- Android emulation. We’ve seen it in Moblin. Will it be a part of the Chrome OS?
- It’s clear now that Android not coming to PC’s through any official channels.
- A Chrome OS using a Moblin 2 core on a Moorestown platform could provide an always-on smartbook experience.
Summary
I think the Google Chrome OS will be a showcase for the launch of Google’s re-written HTML-5 app suite. It will run in an enhanced Chrome envelope and have basic windowing system capabilities. It IS a threat to the low-end of Windows desktop OS where Windows 7 takes 8GB and 1GB of RAM along with a licence fee and it also threatens Moblin but it could be made so thin, resource-light and simple that it could be used alongside existing operating systems using virtualisation techniques. Using the latest low-power platforms for X86 (Intel’s Moorestown) and ARM, the Chrome OS could be used to make stand-alone Chrome hardware devices at very low cost. 3rd-party device support will be a challenge that could be solved by teaming up closely with hardware manufacturers in an expanded version of the Open Handset Alliance.
New article: Google Chrome OS. Round-Up, Podcasts, Thoughts. http://cli.gs/64uu6
One thing that makes X11 look memory hungry is that it reports graphics card memory as memory its using. So if you have a graphics card with say 512MB ram, it will appear as if X11 eats up that much system ram…
Also, the graphics used by all ui elements and similar is stored inside X11 memory, not individual app memory…
Here is what I believe. You have an interesting way to analyse this, I would link to the Masterful John C Dvorak for some very clever guessing: http://www.marketwatch.com/story/is-googles-new-os-more-than-just-a-bluff?siteid=
I do not believe John C Dvorak is 100% right in his funny column, thogh I do believe he is right when he says that this is all a super clever public relations trick put on by Google and that all of it is just the Google OS coming up. John C Dvorak is mostly right about most things that he says.
I believe it will be released open sourced in a couple of months, with the first ARM Cortex A8 and Tegra based laptops.
Android 2.0 and Chrome OS is the same thing. It doesn’t matter what Google says and what bloggers think. There is only one way Google is working towards:
– Making full Chrome browser work on ARM embedded laptops even better than on x86 based laptops.
Now, you might know me as the contiunous x86 basher, I kind of am. But what I believe Google wants is more competition in both hardware and software space for PCs and laptops. This is what Google OS is all about.
The reasons Google might caution Google OS on ARM fans to wait for are a few technological breakthroughs which Google might need before the worldwide availability of perfect $100 Google laptops can happen:
1. ARM Cortex A8 needs to be fast enough for a full browser. If it’s not, then Google needs to wait for broad availability of ARM Cortex A9 starting early next year.
2. Google and the whole ARM community needs to optimize browsers, flash, HTML5 features on DSP and GPU cores of laptops, especially ARM laptops, so that $100 laptops can run a FULL browser and cloud computing experience. Nvidia, Qualcomm, Freescale, Texas Instruments were promising hardware acceleration for the browser, Flash and HTML5 at Computex, but they didn’t really show it yet. I believe they can make it work as a 2003 X86 based browser (something like a 512MB RAM or less system), though that may not be enough for the full mass market to adopt the first version, thus Google might prefer to wait for full launch for it to work better than 2009 x86 browsers.
3. Google wants better connectivity. Google is strongly hoping to start implementing White Spaces worldwide as soon as possible, this will enable free unlimited wireless Internet for all (and destroy all ISPs and telcos in the process). Optimized Connected standby features for ARM devices might only really start working perfectly early next year. First generation ARM Google OS laptops might not have LED lights that turn on instantly on incoming emails, feeds, pings, IMs, VOIP calls and other such crucial presence and social networking web apps which Google needs on the Google laptops for it to really feel like revolutionary products compared to the established systems.
4. Political aspects of this might start being put into places early next year as well such as real competition on HSDPA connectivity, maximum prices of $20 per month pre-paid data-only plans for most of the world and no more contract-plans and other voice and SMS plans forced onto consumers by monopolistic telcos. Also political decision on net neutrality, white spaces, sustainable energy consumption of consumer electronics and servers and crucial for Google to succeed on this global cloud computing plan.
I see it as inevitable, that Google will create Google OS, a super tiny embedded Linux open source OS less than 50 Megabytes for the whole highly optimized OS, and that in a couple of months we will start seeing it ship on $150 ARM based laptops with all types of screen sizes (large screens and keyboards aren’t much more expensive than small ones, consider $50 upgrade for 15″ and full keyboard instead of 10″ and tiny netbook keyboard).
Those $150 Google laptops will be running ARM chips by half a dozen competing ARM processor manufacturers and manufactured by all the major laptop manufacturers in the world. Effectively putting out of business all of Microsoft, Intel and Apple. Together with most of Silicon Valley. That is for the better. For the first time billions more people will have access to this technology very quickly and we will all for the first time really find amazing new ways to use the technology.
As for technical details on Native versus Cloud apps. I believe natively you will have everything needed for a full computing experience. Basically it’s not just the browser, it’s not just flash support, it’s not just HTML5 including native code plugins for the browser and 3D in the browser, it’s like providing you the hypervisors, user interface APIs, clever caching and seamless interface optimizations, which will enable you to not only have a full 2009 x86 style computing experience, it will plug you into the full cloud, in fact giving you infinately more computing power for all the most processor intensive tasks that the biggest professionals would want to use. You can definitely encode videos using grid server encoding, I have been doing that for over 2 years for all my HD video encoding needs, just have a fast enough upload to upload your source files from your camcorders. Google Gears type database and web application caching not only lets you do things while offline, it can turn all web applications into feeling exactly like native applications, they respond instantly without having to wait for any online service to stream the user interfaces back at you. The user interfaces will be locally cached on the machine, only processed data is streamed from the cloud, and clever pre-loading algorithms mostly will not make you feel any difference than processing everything using a local X86 processor. In fact, things will feel much faster cause you will be able to have the power of an unlimited amount of cloud servers to render, process and encode any of your media intensive tasks.
The ‘big brother’ thing should become less of an issue when encryption becomes easier to use. That way your data can be out there but it’s meaningless to anyone who doesn’t have the keys for it.
To see where html 5 support is at:
http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28HTML_5%29
For the mobile OS stuff, I don’t think it will be that way. There is an API that is supported by many web browers called gears, which allowed you to work offline with many RIAs, like Google Docs, gmail, etc.
http://en.wikipedia.org/wiki/Google_gears
The coolest thing I think this OS will do, is if it is adopted by a lot of people, that it will increase the developers creating RIAs and other web based content (more games etc), it will also push HTML 5 support faster, where as for example, it will probably take Microsoft a long time until they support it, and Microsoft strategy is to push silverlight and html 5 would get in the way of thta.
For the location-awareness, Google Gear supports a Geolocation API, which can use any location based tracking (GPS, network/cellid) that is available on a system.
For 3d, I don’t see google making any headway in that direction. First off they are not interested in using graphics hardware acceleration for web pages, they have rather blown it off as not feasible with out having actually tried it out, although since Google Chrome is young and they were just looking into getting out the door faster, maybe in the future they will be interested in it. The other 3d option they wanted to go with is a very high level 3d api, which most high level apis like that are not successful (Java3D, VRML, etc). Mozilla on the other hand has proposed OpenGL ES, and even the Khronos group behind OpenGL has said they want to get 3d to the web, so there is hope still there, just not from google. Apple and Mozilla also have been flirting with 3d CSS transforms, so web page content can be moved and rotated in 3d space (but I don’t even think any of that is official in any standards).
Google Chrome OS. Round-Up, Podcasts, Thoughts. http://bit.ly/QCSBM
“A Chrome OS using a Moblin 2 core on a Moorestown platform could provide an always-on smartbook experience.”
Not sure what is meant by that but if Moblin 2 has a web browser, it will be capable of doing the exact same thing as Chrome OS. As a matter of a fact Chrome OS doesn’t claim to functionally do more then any other OS with a web browser, but rather to provide a very simple OS designed only to provide a web browser, which also has the benefits of better security (less to secure), and faster boot times.
LeeN your point about encryption is wrong in some respects. The problem with encrypting your data then sending it to the cloud is that means it cannot be read in the cloud. Great for security but what if you want to search your data or perform calculations on it without pulling it all off the cloud. This is impossible because the data cannot be read within the cloud as it is encrypted.
This is the kind of chicken and egg scenario which is really going to have to be solved to allow mainstream business to consider cloud based applications seriously.
You can be sure everything is encrypted on Google’s servers, many many times, and they encrypt every aspect of their databases, cache and networks. Only very few, if not only the top of Google has any way to find out those encryption keys for all of Google’s system. If you think any of the 20 thousand employees at Google can read your gmail, you are totally wrong. Google might even have made it so not even Sergei Brin can do anything to try to read your gmail. It could be somekind of random encryption key for each gmail account, stored on random servers in ways that are unhackable.
What Google needs though is higher level of encryption on the client side. I’d like one of those calculator type of things that banks give people, so that I could anytime I use a public terminal such as in a netcafé, I should be able to click a button on Gmail to not only log-off but to reset that extra level of security so next time I log-in I would need to use that unhcackable physical calculator product thing. This way any keyloggers on the netcafé computers would be rendered useless. And any attempt at logging in should be logged by Google and viewable by the user.