It’s a shame but not many website owners or for that matter, web developers are familiar with what SQL Injection is and just why it’s something they need to worry about.  I’m noticing through various forums, friends, etc an increased number of sites being exploited for Cross Site Scripting through SQL Injection.  Most blog readers are going to say “HUH? Cross Eyed Scripting? What Injection?”

Here’s what I’m talking about, a hacker will come to your website and use SQL injection to gain access to your database so he can then run SQL queries against your database directly. He not only will have full read access to your database contents (if it’s an e-commerce site you are encrypting your customer information, right?) but he’ll be able to modify your database and place his own code in there.  What you might find is some iframe HTML code that’s calling some JavaScript file (www.domain.com/hackerscript.js) trying to make that load in your site instead of your actual content. Then some website visitor comes to your site and their browser attempts to run this JavaScript which downloads and installs a Trojan on their computer. Now not only is your site hacked but you’re spreading malware all over the Internet and if you don’t do something about it soon, you’re going to get flagged by Google and the other search engines.

Yes this is a bad bad thing.

UPDATE: Moments after posting this blog post I found a blog post about the United Nations Events page being marked as malware through SQL Injection

Learning more about SQL Injection

I’m not going to rehash all the gory details on SQL Injection but instead recommend you read the following references:

A great article on SQL Injection by Paul Litwin (he’s a programming superhero)

Neil Carpenter’s Blog articles on:  Anatomy of a SQL Injection Incident Part 1, and Part 2,  and SQL Injection a Comment.

ScottGu’s Blog: Tip/Trick: Guard Against SQL Injection Attacks

Michael Sutton’s Blog: How Prevalent are SQL Injection Vulnerabilities?

These articles all have links to other great articles and you’ll learn more than you ever wanted to know about SQL Injection

How do I know if I’m vulnerable to SQL Injection

The problem with SQL Injection is so many people out there are using scripts that they not only don’t know if they’re vulnerable but they don’t know how to fix it. So step #1 is to find out if you’re vulnerable and here’s my recommendations for that:

#1 ScanAlert.com - If you’re hosted at AppliedI.net they’ll provide you a year of quarterly scans for free and then reduced renewal. They also have a service called HackerSafe that does daily scans of your site and I highly recommend you use a service like this as they’ll really dig through your site and alert you that you’re vulnerable before it’s too late.

#2 Secunia.com - They provide a great resource cataloging vulnerable sites and providing frequent updates as new vulnerabilities are discovered. If you’re using any 3rd party applications (freeware, opensource, commercial, anything) I recommend you punch it in and take a look. For a quick reference, here’s a graphic representation of vulnerabilities in OSCommerce

And here’s one for all the Anti-Windows Hosting people.. Here’s all the vulnerabilities found in IIS6 since 2003 to today (there was one released just recently) there’s been 5 and all are either patched or workarounds provided:

Here’s the same graphic for Apache 2.0.X, there’s been 35 and looks like some are still unpatched or only partially fixed:

  From the Secunia Site you can get more information on security holes and the impact they have.  BTW, I did a quick search on advisories for SQL Injection, there were 2072 returned results. Yes, it’s a problem.

#3 Use another security scanner like Nessus, Acunetix, Elanize’s Web Security Scanner, etc.

How do I protect my site from SQL Injection?

So the big question is how do I protect my site from SQL Injection? The answer: YOU FIX YOUR CODE! The problem is not all website operators are website developers and they don’t know how to fix their site or they don’t have the resources (money, experience, etc) to fix their site. What are they to do? THEY HAVE TO FIX THEIR CODE! If you can’t fix it, I recommend you find a replacement application that’s not susceptible for SQL Injection. If you can’t find a replacement application and can’t fix your code here’s a few tips that might help but this really needs to be address at the application itself:

IIS 6 SQL Injection Sanitation ISAPI Wildcard - It’s an ISAPI filter that intercepts all requests and cleans out the dirty SQL injection code.

Port 80’s Server Defender - It protects against a lot more than just SQL injection and is billed as a Web application firewall.

A little ASP Script for cleaning inputs that you can incorporate into your application.

A little guidance from MSDN.

And numerous other resources out there for your review.

Conclusion

I don’t know why I always have a conclusion but you have to close your article somehow. In conclusion, I hope you spend some time and read a few of the articles linked to in this blog post. You’ll learn more about SQL Injection and hopefully avoid being compromised by it.

A question that’s been coming up rather often lately is: “How do I configure the Windows Firewall on my Dedicated Server / VPS Server?” Not many people realize that in SP1 Windows 2003 got a software firewall feature added and even fewer are familiar with how to configure it correctly.  Because of this I wanted to provide a few tips on how to configure your Windows Firewall.  Please note that this is NOT the all encompassing tutorial on how to secure your server or how to use Windows Firewall but it’s a great starting point.

Introducing the Windows Server 2003 Firewall

Many people are aware that Windows Server 2003 has a GUI for the firewall. You can access it from the start menu as shown below:

When you first click that icon it’s going to enable Windows firewall and present you with a nice GUI and if you’re not careful you’ll proceed to get yourself in a lot of trouble!

I really don’t advise you configure Windows Firewall from the GUI, atleast initially. The reason for this is if you go to exceptions you’ll only get a few options:

Sure you can enable an Exception for Remote Desktop but what about IIS, DNS, maybe Mail? In fact, the most common support question regarding Windows Firewall is “I turned on Windows Firewall and now I can’t connect to my server via Remote Desktop any more, HELP!”  It happens

Thus enters the Command Line Interface (or CLI as we like to call it).

Managing the Windows Firewall via CLI

The real beauty of configuring the Windows Firewall via CLI is that you have much more control over it. In fact if you open a dos window and type:  netsh firewall set portopening
you’ll get output like this:

Which provides all the gory details on how to configure the Windows Firewall port openings.  A few things about the Windows Firewall in Windows Server 2003,

1. you can only block inbound packets (you have no control over outbound packets, that is available in Windows Server 2008 though).

2. In a normal firewall you can block a port for specific IPs. So let’s say you’re getting attacked on port 80 from 10.5.5.5 you can block that IP on just that port. You can’t do this with Windows Firewall In 2003.  So with Windows Server 2003’s firewall you can do the following with a port:

1. a. block it for everyone from the outside
2. b. allow it open for everyone from the outside.
3. c. allow it open for only certain IP addresses.

That’s the bad news, that’s really all we have to work with, the good news is in Server 2008 so much more is available and I’ll blog about that later.

If you want a full list of what you can do, you can go to this article on technet:

Windows Firewall Tools and Settings (server 2003)
Windows Firewall Tools and Settings (server 2008)

The AppliedI Quickie Firewall Ruleset

In an earlier blog post on the AppliedI blog, I spoke about Joe Healy and mentioned a little cookbook of firewall rules I like to enable on my own personal VPS servers. These rules are here:

   1:    netsh firewall set opmode enable 

   2:    netsh firewall set portopening TCP 80 HTTP 

   3:    netsh firewall set portopening TCP 53 DNS-TCP 

   4:    netsh firewall set portopening UDP 53 DNS-UDP 

   5:    netsh firewall set portopening TCP 21 FTP-Server 

   6:    netsh firewall set portopening TCP 220 IMAP3 

   7:    netsh firewall set portopening TCP 143 IMAP4 

   8:    netsh firewall set portopening TCP 25 SMTP 

   9:    netsh firewall set portopening TCP 110 POP3 

  10:    netsh firewall set portopening TCP 3389 RDP 

  11:    netsh firewall set portopening TCP 443 HTTPS 

  12:    netsh firewall set portopening TCP 9999 SmarterMail 

  13:    netsh firewall set portopening TCP 9998 SmarterStats 

  14:    netsh firewall set logging droppedpackets=enable

Line #1 turns on the firewall. Lines #2-13 enable various port openings for all IP Addresses and Line #14 enables logging.

The only caveat I’ve run into this set of rules is that passive FTP no longer works as a result of this change. The workaround for this is to enable Passive FTP on a given set of ports and then open these ports in the firewall as well. Here’s a KB article on how to do this. You’ll also want to open these 100+ ports on the firewall and here’s a little code (from here) on how to do that:

To add a range of ports to Windows Firewall from the Command Line

1. Click Start, click Run, type cmd, and then click OK.
2. Type in the following where the range is specified in ( ) and the name of the firewall entry is in ” “.
   FOR /L %I IN (5001,1,5201) DO netsh firewall add portopening TCP %I “Passive FTP”%I
3. Each port in the range will be added with an “OK” confirmation.

This opens more ports than you’ll probably need and you can adjust these accordingly.

Opening a Port for a specific IP Address

A fairly common request we get is how do I open a port for only specific IP addresses and block it for everything else. This comes when a user wants to open port 1433 to their SQL Service so they are able to access it remotely using SQL Management Studio but leave it closed to the rest of the world. 

Let’s assume we want to open PORT 1433 for TCP requests only and only to the IP address 10.5.5.5 the way to do that is with this CLI command:

   1:  netsh firewall set portopening protocol=TCP port=1433 name=MySQLAccess mode=ENABLE scope=CUSTOM addresses=10.5.5.5

This opens that port for a custom scope of addresses which in this case is only 10.5.5.5

What else can I do?

Well, there’s a lot more you can do with Windows Server 2003’s Firewall and this really just scratches the surface on what can be configured and how. Many people complain a software firewall is not a good firewalling solution because if a hacker can get into your server they can disable it or do what they wish to it. But I believe the added layer of security a software firewall brings to the table in addition to other network security features in place (such as network firewalls, IPS, IDS, etc) that it’s only wise to run it as well.

Ever since Bill Staples first blogged about the collaborative work Microsoft was doing with Zend to improve PHP stability and performance on IIS, I’ve been giddy as a school girl. Having run a core component of the AppliedI.net infrastructure that makes use of PHP and MySQL on IIS7 and making use of FastCGI and the new Windows CGI friendly DLLs/builds it’s made me even more giddy, PHP simply zips along these days on our windows servers and I can’t wait to move my blog to a box running FastCGI.  This week at ZendCon Bill announced some more good news: SQL Server 2005 Driver for PHP.

From the Microsoft Data Blog:

The PHP Driver will deliver an API designed-to-enable reliable, scalable integration with SQL Server for PHP applications deployed on the Windows platform. Zend and Microsoft are working together with the PHP Community to ensure that the SQL Server 2005 Driver for PHP is a great offering for PHP developers and part of Zend’s core PHP offering.

So that’s the good news, the bad news is we have to wait one more day, because the download won’t be available until October 11th at: http://www.microsoft.com/sql/technologies/php/default.mspx

I’d be remiss if I didn’t point out here that I was recently interviewed by HostSearch.com.  HostSearch is an online hosting magazine that covers just about everything hosting related. You can see the Interview here.

If you’re seeing viewstate errors like “The viewstate is invalid for this page and might be corrupted” here’s the dealio.  For security ASP.NET encrypts the viewstate using an Autogenerated Key that is generated when your application pool (or worker process) is started. At AppliedI.net (and many hosts today) each website is placed in a unique application pool so your site is isolated from the other sites on the server. If your application pool recycles for whatever reason, the viewstate key will change and when you go to post back to your application is may fail with the gloriously informative error of:

“The viewstate is invalid for this page and might be corrupted”.

What’s causing the application pool to do this?

Your application pool is recycling due most likely to a couple things:

1. you edited your web.config and that caused the change (not likely)
2. IIS was reset on the server and that caused the change (less likely)
3. Your application errored out causing the application pool to recycle (possible)
4. Your application pool reached a memory limit and was forced to recycle (I put my money on the #4 horse to win! this is probably it)

All hosts today that use dedicated application pools, isolated application pools or “website sandboxing” whatever they may called also set a memory limit on the application pool in shared hosting, some set it more aggressively than others.  The advanced web applications of today are using more and more resources as they are more and more complex, it’s not uncommon to find a web application using anywhere from 150-350MB of memory today, although the average is still just in the range of 80-120MB. 

The good news is you have options on how to get around this. 

1. You can upgrade your shared hosting account to an account that has a higher memory limit. At appliedi.net we offer 4 different shared hosting accounts each with separate memory limits.
2. You can upgrade to a VPS hosting account or dedicated hosting account and set an even higher memory limit.
3. You can completely disable the viewstatemac by adding “enableViewStateMac=”false” in your web.config. This would be a bad thing though and you can google viewstate injection for why this is bad.
4. You can generate a predefined key and stop using the default autogenerated key method.

#4 is probably the route you’ll want to go. My recommendation is to go with #1 first and not just because I work for a hosting company and want to see you spend more money per month. I recommend you upgrade your hosting account to an account that has a reasonable memory limit for your application because every time that application pool recycles your site is going to be slow as it recompiles and initializes the application.  You also just lost your session state information so if you’re an ecommerce site and your client was in the middle of a checkout, they probably just lost their cart contents and have to start over.

Creating a predefined viewstatemac key

The guys over at www.aspnetresources.com have a keycreator tool that makes generating your own predefined viewstatemac key effortless. To do this you’d do the following

1. visit: http://www.aspnetresources.com/tools/keycreator.aspx and use it to build a key
2. Copy this key into your applications web.config file between <system.web> and </system.web>
3. Save your web.config file and test your application to make sure it still works. If for some reason it doesn’t work, you can delete the changes and revert back.

Where to learn more

There’s always google. But Microsoft has a pretty extensive article in their KB at: http://support.microsoft.com/default.aspx?scid=kb;EN-US;829743 that will also help. If this sounds like something that’s happening to you at AppliedI.net on your ASP.NET hosting account please contact our support team and they’ll be happy to help you troubleshoot the issue.