Slashdot Mirror
Microsoft Attempts to Secure IIS
billmaly writes: "Yahoo has this article about trying to make IIS more secure. Among steps is to have it install in its most secure state, putting the onus on sysadmins to remove it from that state. It looks like Microsoft may be trying to do the right thing from a security standpoint, at least on paper."
because 78,417 Nimda hits are more than enough for me!
Admittedly, IIS does run certain scripts and perform certain functions as a "nobody" user. But most of the recent exploits were able to get an immediate "root shell" because the services being exploited did run as SYSTEM. And unless Microsoft is willing to address that problem, admins who need to enable many services and don't keep up on patches will still get rooted on a regular basis.
-sting3r
Open the source. Put it up for peer review. Fix the holes. I'm not saying that they should hand out the source for their whole OS, but when they have had as many debacles with one piece of software it might actually help them out quite a bit.
I refuse to install products that require IIS as well. A software provider of ours makes an ultra nice business mining product that can be nicely web enabled. I told them that I would purchase it as soon as they supported a web server that didn't have a new security flaw or bug discovered every week.
I had to test some java code being developed by (company) for a newly released (product) and needed a web server. The usual test platform server had just been taken down by nimda (ie not 3 hours earlier). Fortunately for my productivity log, an extremely capable app called Apache exists for WinNT and in under 30 minutes I had it up and running (including denying every host under the sun that was sending those annoying GET requests for /winnt/system32/cmd.exe).
:-)
The entire dev team working on the java code would have just taken the afternoon off, had I not casually mentioned the existance of my humble Pentium Pro 200 running Apache.
This caught the attention of my boss who wondered why our group was able to continue working, while many others were outside playing basketball waiting for the Admins to finish the virus updates. Who knows . . . we may shift away from simple IIS servers (for a java service on a server you don't need some big IIS machine).
From a security stand point, This little server did a good job of fending off every virus attack (a few hundred every hour). I believe two additional simple IIS servers have been temporarily changed to Apache since they don't have a need for any other service. Who knows what will be their ultimate fate. But right now they are doing their job and don't need to be updated. This may affect the purchasing policy for one or two machines here. Not a huge step towards non-M$ product use, but I am encouraged none the less.
robi
The rest of this comment is from the NTBugTraq newsgroup:
Microsoft have today announced a suite of initiatives intended to address the issues their customers face from the threat of Worms and other malcode like Nimda and Code Red.
About time.
I've been assured that substantial resources have been allocated to this new effort, but one has to wonder just who was consulted in coming up with what this program involves (if you were, drop me a line.)
Announced today was the "Microsoft Security Tool Kit";
Click here
This "Greatest Hits" CD or network download contains all of the things you should already have;
- - Latest Service Packs for OS, IIS, and IE.
- - Security Checklists for NT, W2K, and IIS.
- - A W2K-SP2 Deployment guide (the Update.msi section is worth reading if you have an Active Directory environment and use Group Policies)
- - An NT 4.0-SP6a Deployment guide for SMS.
- - IE Deployment guides.
- - Several individual Hotfixes required for NT 4.0 Terminal Server (even though they are included in the NT 4.0 SRP) - - IIS Lockdown Tool
- - URLScan
- - HFNetchk
- - Critical Update Notification 3.0 (only applies to W98/W2K according to the referenced KB article)
- - QChain
There's a difference between the download and the CD. According to the announcement page, "It (CD) includes automation scripts to quickly install all the security hotfixes recommended in the kit.", but the CD may take from 3 to 6 weeks to arrive.
I was told there would also be a "Bootstrap Client for Windows Update" within this package somewhere, but if its just the Critical Update Notification 3.0 tool then its not a "Bootstrap Client" in the sense I thought it was.
While there are additional things planned, the biggest thing missing at this stage is a re-release of the NT 4.0 Option Kit CD which contains;
1. Patched version of IIS 4.0 (one that's not vulnerable out of the box)
2. Patched versions of MDAC
3. Modifications to the samples to eliminate RDS
4. Modified default installation that doesn't install in a way known to be exploitable
5. Modified Setup program that doesn't re-install removed script mappings and other components after the user has manually removed them (since that's what many people have done to protect themselves)
In addition, what is desperately needed is some way to do the following;
a) Probe your internal network to identify IIS installations (this can be done with HFNetchk, but working with its output is no fun) /scripts, tightening
permissions, etc...
b) Completely remove the IIS installation on command (remotely!), or render it stopped
c) Query the IIS installation and alter it, removing RDS keys, updating MDAC, patching it, disabling
d) Report results in a comprehensive fashion
I don't know about the rest of you, but many people have thousands of IIS boxes to deal with. While Microsoft does sell SMS, if you used Ghost to distribute your installations it hardly seems reasonable for MS to expect you to purchase SMS to secure what you thought was a reasonable installation.
If you have more than 1000 hosts under your control, send me your suggestions for the best product/method used to get patches and service packs out.
Given that this whole initiative, supported at the highest levels in Microsoft, is designed in response to Worms that required the touching of every machine in your organization, the first thing out the door should've been something that made that problem less onerous.
There are plans in the works (for Q2-2002) for an internal version of Windows Update. I've been calling for this with Microsoft for eons now, and while its great they have finally been hit with the clue-bat it seems ridiculous that its going to be 6 months plus before we see it. Such a tool would allow Network Administrators to rely on the client's Windows Update component to provide fixes (fixes decided on by the Network Administrator). In addition, a new feature in that client (still some 3 months out) allowing it to be setup to allow automatic updates (a push mechanism), would give you a way to push out a fix quickly to all clients.
Again, about time!
Also coming out of all of this was news that Windows 2000 SP3 is not likely to ship this year.
Cheers, Russ - Surgeon General of TruSecure Corporation/NTBugtraq Editor
"I thought I had an Appetite for Destruction, when all I really wanted was a club sandwich."
These are the guys who have still be unable to figure out that the Buffer Overflow, etc. patches are available to them on Windows Update--or that almost all the new exploits would be fixed by getting Service Pack 2.
Um, I think you've completely missed the point. First off, not all patches are available from WindowsUpdate. In fact, precious few are. Most of the updates from WindowsUpdate apply to IE, not IIS. Second, there are a large number of exploits that have appeared since SP2 shipped. I have personally installed nearly two dozen Post-SP2 hotfixes to one server. I average between 8 and 10 post-SP2 hotfixes per server.
Mind you, actually keeping up-to-date on hotfixes actually became possible with the release of HFNETCHK. Before then, it was virtually impossible for any normal sysadmin to keep up with all of Microsoft's patches and apply only the ones they were supposed to. Also, before the release of QCHAIN, it was a horrible and time-consuming process to apply hotfixes to a server, even when you knew which ones to apply, because each hotfix wanted its own reboot to complete and you couldn't just apply them all and then reboot once.
I actually use WindowsUpdate, HFNETCHK and MPSA to check and make sure I catch all possible vulnerabilities. I've found that it's not uncommon for each one to catch something the others did not.
Even with the three tools I listed above, properly securing IIS (or any MS server) is still a royal pain. The damn things come preconfigured with their flies completely unzipped. MS's IIS Lockdown Tool won't even run if you've already taken some steps on your own to manually lock down IIS, and even if it does run, it doesn't turn off the "../" parent directory functionality that's enabled by default. You still have to go into IIS Admin and turn that damn thing off manually.
Let's not pick on IIS admins unfairly. Many of them prefer Linux and use it at home, but have to use IIS at work because that's been mandated. Debian makes it easy to stay patched and does a half decent job of implementing default security, but MS leaves everything wide open by default, makes it damn difficult to lock any system down effectively, installs unnecessary services by default (and won't even let you uninstall some of them) and has a half-assed mechanism for rolling hotfixes and patches out to customers.
Microsoft needs something like Symantec's LiveUpdate, which allows sysadmins to roll out tested updates to internal users on their own schedules, without physically touching every system on their networks. Yes, there are IIS admins out there who are jackasses, but there are plenty of overworked sysadmins out there who'd love to properly secure IIS, if only it weren't damn near impossible.
In other news, Microsoft's hardware division announced a plan to make water flow uphill.
Thank God. Since MS usually tries to do the wrong thing, on purpose. Now they are doing the right thing on paper.
In other news today, Satan said to be interested in joining US Figure Skating Team. "Yes, this is a serious bid; we've already started training now!", said the Dark One, executing a perfect double axel over what was once the Ninth Plane of Hell.
Personally, I would think that rewriting from scratch would make IIS more dangerous. At least Microsoft is plugging the security holes. I would think that rewriting it from scratch might cause more new exploits, whereas fixing the old version makes it more secure with every revision.
This article, on the other hand, shows that Microsoft is trying hard to actually make its product better, instead of just saying "Here it is. New version. Use it or be forever left behind..." like they did with Office XP. I think this goes to show what a company in a non-monopoly position will do to succeed. (No one has a monopoly in web servers, and Microsoft isn't even the leader...)
This is a good thing, and it's the right choice for Microsoft. Please don't call for a rewrite, or in two years we'll all be complaining about the root exploits discovered in the new IIS...
"You are running Outlook 97 or Outlook 98. You should consider upgrading to the latest version of Outlook to ensure you have the most recent product and security enhancements."
Hmm. Is this telling me that there are no patches available, and my only choice is to pay cash money and upgrade to Outlook 2000?
Yeah, it provides useful information, but it still feels like they're trying to shaft me.
-grendel drago
Laws do not persuade just because they threaten. --Seneca
While I easily see your point, it doesn't solve the fact that most IIS admins are complete morons for leaving the systems unpatched to this point.
My point about Windows Update is that ALL of these recent high-prifile attacks have had Windows Update patches for MONTHS. Service Pack 2 blocks almost all of them as well.
I have seen entire tech department that were knocked out by Code Red. Then Code Red II. Then Nimda. Yet, as a "casual" IIS user, I was never hit AT ALL. These patches have been obviously available for MONTHS. And even after Code Red, IIS admins STILL couldn't figure out to patch a hole that has about 4 OBVIOUS places to get the patch from. Let's review.
1) Windows Update
2) Service Pack 2
3) MPSA
4) Any of the virus scanner's homepages which linked to patches after Code Red, Code Red II, and Nimda.
If IIS admins can't even patch the obvious stuff like that, there is really little hope.
As you say "Many of them prefer Linux and use it at home, but have to use IIS at work because that's been mandated."...they are the PROBLEM, not Microsoft. HFNETCHK is easily available, and if Linux users are too lazy to learn how to admin the system that they're PAID to admin, they deserve what they get. I don't care if you don't like Windows, if it's YOUR JOB to be a IIS admin, you sure as heck better learn how to do it RIGHT.
I'm sure modders are gonna hate me for saying that, but I don't care at all if you don't like the system. If it's your job, it's your job. I hate Oracle, but that doesn't mean I don't use it *right* when I have to. Is it my first choice? No. Am I gonna be a slack-ass about it just because of sour grapes if I have to you it? No.
-Jayde
P.S. Disabling Parent Paths is not a big deal if you secure the rest of you system. In fact, I doubt you would find any professional IIS web server which has Parent Paths disabled, as it has terrible effects on most ASP code. It's stupid for server-side code to be forced to code paths based on the root "./" instead of relitive paths "../" as server directory structure could easily change at any time.
What's a sig?
Microsoft has done an excellent job at reducing the amount of excess water in their certifications with the new rounds of exams. I've taken and passed my Windows 2000 MCSE (after my Solaris SCNA and Cisco CCNA certifications) and I found the design exams to be especially challenging. To pass the new Windows 2000 tests, you MUST have experience with deploying their products or you WILL fail.
Cut Microsoft some slack in their certification department. They've came a long way in establishing a well-known industry standard and now they are "fine-tuning" it to ensure that its worth stays intact. As someone who has gone through the process, it holds a lot of value to myself and my clients and customers.
On a side note, pick up a Solaris book at Barnes and Noble and read it for two days. You can pass it without almost no experience, other then knowing run levels and where rc files are located.
-Pat
I would think that rewriting it from scratch might cause more new exploits
Yes. That's why sendmail and bind are the paragons of security they are today. From-scratch attempts to replace them are riddled with holes that make IIS look like a pinprick.
In the spirit of hfnetxchk.exe there is now a tool to apply multiple hotfixes without rebooting, qchain.exe
To use this, you write a
Or at least that's the theory. The hotfixes I was working with didn't all honor the "no reboot" switch. I don't have the list handy (I've since been laid off and don't have access to the network directory with the
The really keen thing to do, for desktops anyway, is to use hfnetchk to identify machines needing hotfixes, a script to customize the
Veteran, Bermuda Triangle Expeditionary Force, 1992-1951
When a (h)(cr)acker writes a virus/worm that cracks into servers and provides root access without actually doing any damage, what they are doing is letting the world know how easy it is to do so.
Bear in mind that there are lots of folks out there (thieves, terrorists, enemy governments) who would (and presumably do) break into servers and steal credit card numbers and/or sensitive corporate/government info, without telling anyone!!
If the "virus authors" weren't constantly exploiting these simple security holes, the greater public would never know they were there, because the real "bad guys" always try to go unnoticed.
I wanted to post this but you were ahead of me. And it's not just a problem with IIS -- most (all?) NT "services" run as LocalSystem, which actually has even more privileges than Administrator.
Bugs and security holes are inevitable in any software, but their impact is different. Any buffer overflow in IIS is disasterous, whereas a buffer overflow in Apache will have a very limited damage. To 0wn a Unix box running Apache you need two security holes: first a hole in Apache to get unprivileged access, then another hole elsewhere that lets you get root. This is considerably harder and a lot more unlikely than a simple buffer overflow in the web server.
On top of that there is a huge problem with file system permissions. Both Unix and NT have the ability to restrict access to files. The difference is that a default installation of NT has all file permissions set to Everyone:Full Control(*). (That's like making every file and directory 777)! You have to manually lock it down! If the file system permissions are not used, running IIS as an unprivileged user won't help.
Contrast this with Unix. Even if a hole in Apache is exploited, you won't even be able to overwrite the web pages (unless another hole is used to gain root access, see above).
(*) I understand the default file permissions have been improved somewhat in windows 2000. Could somebody in the know give more details? Oh, and what's the deal with IIS running partially in the kernel? is it true or has it been debunked?
In all fairness, Unix has had its problems with root-running daemons. BIND was the latest exploit. Since then BIND guys have learned their lesson -- version 9 no longer runs as root. Will Microsoft learn? After so many years of beeing plagued with security holes, not bloody likely.
___
If you think big enough, you'll never have to do it.
The problem is that parent paths aren't automatically blocked from going any higher than \InetPub\Webroot, which to me is a huge security hole. Yes, properly-secured NTFS ACLs on the filesystem will prevent any real damage from occurring, but NT and Win2k default to EVERYONE|Full Control on all filesystems, both at the NTFS ACL level and at the share level.
Look, if it were possible to just fix your server once and then not have to go back and fix the same flaw again (and again and again...), more NT systems would be properly patched, but Microsoft seems to have gone out of its way to hose NT 4.0 customers. Win2k does finally let you patch your install folders with updates from the service packs, but NT doesn't let you do that, and there's no good reason for that. Any time you add or remove a service in NT, you end up putting the install CD in. The second you do that, you have to re-run your service pack and reapply all of your hotfixes.
IIS 4.0 is the current version of IIS for NT 4.0. Let's say you decide you want to build an Outlook Web Access server for your organization and your company hasn't moved to Win2k Server yet, so you use NT 4.0. How do you get IIS 4.0 on that server? You use the Microsoft Option Pack 1 for NT 4.0. Guess what? That thing installs an insecure version of MDAC, an unpatched version of IIS and a host of other crap you may or may not want (such as the MS transaction server and indexing). All of it is incredibly old and almost all of it has to be patched and repatched the second you install it.
So, here's how you build your server: You install NT 4.0 and apply the latest service pack (SP6a because SP6 had heinous bugs). You install IE 4.0 or newer. Then, if you're smart, you install a version of MDAC (2.5 or newer) that sets proper registry security and is reasonably recent and free of its own security holes. Then you install the Option Pack so you can have IIS 4.0 and which insists on trying to install MDAC 1.5--be sure to deselect RDS because that's a huge security hole that Russian hackers use to steal credit card numbers. Now, you're ready to install Outlook Web Access. Think you're finished? Ha! Not even close. Next, you run HFNETCHK to find the enormous list of hotfixes you've got to download and apply. Each hotfix is in a different place on Microsoft's website, and there isn't a convenient tool you can use to just go and download the patches you need and store them in conveniently-labeled folders. Then, you download QCHAIN so you can apply those patches without having to reboot after each one. If you're smart, you'll use WindowsUpdate and MPSA to make sure you're not missing anything.
By the time you've finished with this minimum effort, you've spent no less than four or five hours just installing NT, IIS and the hotfixes, not to mention the hour or two it takes to install and configure OWA. Now, at this point, all you have is a product that's reasonably free of serious buffer overflow security flaws. You still don't have a product that's actually remotely secure. Now, you have to go and fix all of MS's idiotically optimistic NTFS permissions and find and disable any unnecessary services. Maybe you run MS's IIS Lockdown tool, which removes the IISamples folder and a few other obvious things.
By now, you've probably spent at least 8-12 hours building this server, patching the holes and fixing the default security settings.
So, you've patched the living hell out of the server and it's ready to go. You're immune to attacks, right? Almost certainly not. New holes are found in IIS every week and keeping on top of them is a huge job even if you have no other job responsibilities. Add to that the fact that any time somebody adds or removes a service from NT, you have to reapply the latest service pack and all the hotfixes (in order) and then reboot, and you've got yourself a nightmare.
Let me be clear.
There are enormous numbers of jackasses running IIS who can't figure out how to toast bread. However, there are plenty of overworked sysadmins who're only trying to keep their damn networks running who find it nearly impossible to keep their IIS servers patched and locked down because Microsoft makes it so damn difficult.
Yes, matters get a little better when you're running Windows 2k server, but things don't turn into a panacea just because you can patch your install media and some hotfixes don't require reboots. Microsoft still releases at least two or three patches for Win2k and/or IIS every month (sometimes they release that many in a week). They still automatically set file and share privileges too optimistically. They still install dozens of unnecessary services by default. They still force you to have unnecessary applications installed by default that you can't remove without pliers and a blowtorch (OutlookExpress). In short, they still don't take security maintenance seriously and until they do, it'll be tough for even conscientious admins to keep up. Newbies, idiots and lazy bastards won't have a hope.
I have personally seen service patches and hot fixes blue screen servers. I have a fear of installing Microsoft "fixes" on systems that are functioning - will they cause a blue screen when the inevitable reboot is required? Will they break an API my "turnkey" vendor relied on?
I have two choices:
I can pro-actively install the service packs and hot fixes, causing (at best) some downtime or (at worst) an extended period of downtime thanks to unexpected side effects. If I am pro-active about fixes, I am viewed by departmental managers and users outside of IT as a bad guy, someone who is here to wreck their server. Oh, and don't tell me to test it before I apply it... you can install the same service pack on 50 boxes and only have it blue screen on one. I've SEEN this occur, so it is always a roll of the dice.
Choice #2 is to wait until the virus/trojan/whatever hits this department. Then I am the good guy for coming to the rescue.
What would YOU do?! I'd especially like to hear from seasoned sysadmins in both Microsoft and Unix camps - what approach do you take?
-hj