Google Reportedly Ditching Windows
Reader awyeah notes a Financial Times report that Google is ditching the use of Windows internally. Some blogs have picked up the FT piece but so far there isn't any other independent reporting of the claim, which is based on comments from anonymous Googlers. One indication of possibly hasty reporting is the note that Google "employs more than 10,000 workers internationally," whereas it's easy enough to find official word that the total exceeds 20,000. "The directive to move to other operating systems began in earnest in January, after Google's Chinese operations were hacked, and could effectively end the use of Windows at Google. ... 'We're not doing any more Windows. It is a security effort,' said one Google employee. ... New hires are now given the option of using Apple's Mac computers or PCs running the Linux operating system. 'Linux is open source and we feel good about it,' said one employee. 'Microsoft we don't feel so good about.' ... Employees wanting to stay on Windows required clearance from 'quite senior levels,' one employee said. 'Getting a new Windows machine now requires CIO approval,' said another employee."
.....if Microsoft employees can ditch Google.
That will be the true test of Google's influence.
I recently left IBM, but while I was there, there was considerable effort to eliminate M$ products. Symphony was being pushed out over MS Office, and Apple netbooks were an available option in some areas. Obviously IBM has a love for Linux, and the Linux folk there are doing everything they can to make it perfectly acceptable, and usable, to use Linux internally. For all of my 4 years at IBM I used Debian and then Ubuntu on my work thinkpad (but I kept a XP partition for Visio).
Well, I think they're headed to ChromeOS long-term. While this particular report may be true or not since it's based on anonymous sources, Eric Schmidt himself said that this would be Google's response during the Atmosphere event. He also indicated that they're moving toward eating their own dog food at every level, and that wasin or around a discussion of ChromeOS (I'm going from memory). I took the interview as a whole to be an indication that Google wanted to move to ChromeOS and Apps for as much of the internal stuff as it could.
Here is a report of the interview: http://news.cnet.com/8301-30684_3-20002315-265.html
Put identity in the browser.
So Google employees don't use the client software they themselves produce, considering that a lot of it is still Windows-only?
I would be particularly curious about Google's own GTalk client...
It has always surprised me how few companies run linux on the desktop. I have personal converted about 30 in the last 10 years, all of which were mom and pop places with less than 100 seats. Google using Chrome would not surprise me. 90% of the office desktop users dont need more than a browser, office platform, and maybe e-mail assuming the company does not have a web based e-mail. I have heard many geeks say it is not ready for the desktop based on a list of reasons but the general office user has such a small software need that it fits nicely..
The last company I migrated over to linux was a rush job. They needed it done in a short window before the inspection of there licences. I set up 1 server with home directory shares in both NFS and Samba, ldap, dns, printers, and DHCP. There were 3 desktop configs, 1) for users that had with firefox, OpenOffice, and google chat. 2) for managers that had that plus planner, and Dia. 3) was for upper management that had everything from the first two plus a few specialized things that one VP seemed to think he needed like bit torrent and an RSS feed reader.
Everyone got the basics like a calculator, archive manager, Notepad, etc.
All in all they run smooth, easy access to pen drives etc. Windows Laptops could be pointed at the server and after logging in would get the users home directory allowing them to easily move data between there laptop and the desktop. The remote home directories and ldap logins meant that users could login at any desktop and do there work. All the desktops were the same for a given group so if one failed it was simply replaced and a new image installed (Totalling about 45 min install time) Top this off with no viruses, spy ware, or bot software and the desktops were locked down with only a couple of open ports. So far every company I have done this for has loved the setup.
Really? I find it EXTREMELY useful in the music studio. I guess I missed the memo that this is not supposed to be the case.
Information wants to be anthropomorphized.
Or the denial of service attacks possible against an Internet-exposed Exchange server because it simply cannot handle a reasonable amount of direct SMTP traffic, especially broadly distributed spambots?
That is so true. Our Exchange server was falling over at least a couple times a week, even though it was on a fresh install on good hardware and run by a competent admin. It just couldn't stand up to all the dictionary attacks and other jackassery thrown at it. I installed a FreeBSD+Postfix server in front of the Exchange server and configured it to learn which usernames were valid on the Exchange, set up Spamassassin, and let it go. We literally haven't had a single unplanned outage on Exchange since that day.
Dewey, what part of this looks like authorities should be involved?
Personally I've found Pages/Numbers and OO//Calc to be pretty terrible compared to MS Office. That said, at the last company I was at our Receptionist / Office Manager used a Mac and had no problems. So its doable.
I wonder what Google uses for an accounting package? Very hard to find accounting programs that do not require Windows OS.
Corporate accounting? General ledger, accounts payable, that sort of thing? No company of Google's size would do that with a Windows-based application. They would likely do accounting with SAP or Oracle, probably running in a Unix environment of some kind. Both of these have web UIs nowadays, so all the employees who need access can use any OS they want.
Unpatched 5% (11 of 217 Secunia advisories)
That's the important part. Linux always has more vulnerabilities publicly found and fixed due to it being open source, a process which leads to a more secure system -- wouldn't you rather have a vulnerability found and fixed, or even found and marked "unpatched" on Securina, than found and exploited (hidden) elsewhere?
And even more important is what those unpatched vulnerabilities actually are:
A vulnerability has been reported in the Linux Kernel, which can be exploited by malicious, local users to cause a DoS (Denial of Service).
This is in the CIFS code, which presumably can be disabled. Should be fixed, but how many Linux systems actually need to defend themselves against local DoS attacks?
Tony Griffiths has reported a vulnerability in the Linux Kernel, which can be exploited malicious, local users to cause a DoS (Denial of Service).
Another local DoS. And another, and another... Yawn. Let's skip to the good stuff:
An error in the DRM (Direct Rendering Manager) drivers due to insufficient DMA lock checking can be exploited to crash the X server or modify video output.
Modifying video output could be very bad, but also very hard to exploit in a way to make it worse than rickrolling you. And again, local.
A race condition within the handling of "/proc/.../cmdline" may disclose the content of environment variables of spawning processes.
In other words, there's a race condition (hard to exploit) which may disclose sensitive information in your environment variables to other procesess you run. I honestly can't think of a single case where this would reveal anything exploitable. Clearly, it should be fixed, but right now, you're welcome to my environment variables.
A race condition within the memory management can be exploited to disclose the content of random physical memory pages.
That could be very, very bad, but also very difficult to exploit. Again, local.
The vulnerability is caused due to an unspecified error within the ide-cd SG_IO functionality. This allows a user with read-only access to bypass these permissions and perform write and erase operations on media in a drive.
So, in other words, anyone who uses an IDE CD-RW drive is vulnerable. Otherwise, you need a lightning-quick exploit to grab someone's blank media and burn something evil to it. I'm quaking in my boots.
The problem is caused due to signedness errors which can lead to integer overflows in the XDR decode functions in kNFSd. This can be exploited by sending packets with a write request larger than 2^31, causing the system to crash.
In other words, doesn't affect people who don't run NFS, or specifically kernel NFS (there's a userland NFS now). Oh, and you need to be on the local network.
Various functions in the IEEE 1394 driver contain integer overflows within the memory allocation scheme. This can potentially be exploited via specially crafted requests, which may cause a large amount of data to be copied into an insufficiently sized buffer.
That's probably the most serious one I've seen -- possible privilege escalation -- but what privileges do I have to have to access the raw FireWire device anyway? I bet most users can't.
So that brings it down to, what, one actually unpatched vulnerability that I'd be worried about. And it's still only local, and still a bitch to exploit.
Now let's try the Windows ones. One is a remote exploit, which can be triggered merely by convincing an Aero user to view a given image. Another is a remote exploit which may allow people to manipulate SSL-encrypted streams.
Security is not and never has been about numbers -- I only need one serious exploit.
Also worth
Don't thank God, thank a doctor!
When I talked to a Google recruiter a couple of years ago, they said that although you had a choice between Linux, OSX, and Windows, you would have a hard time as an engineer if you used Windows, as about the only people who used it were managers that were running Microsoft Project.
Why can't I mod "-1 Idiot"?