So Windows Update is a dog, now what?

From: Russ (Russ.Cooper@RC.ON.CA)
Date: 04/15/02

Date:         Mon, 15 Apr 2002 13:18:46 -0400
From: Russ <Russ.Cooper@RC.ON.CA>

Ok, I think the previous message about Windows Update should've
sufficiently convinced you that WU is a dog and shouldn't be used, by
anyone, at any time, for any reason. It just sucks, period. And by
extension, that means that the soon-to-be-released Windows Update,
Corporate Edition, should be put into the same boat.

So what now?

How about someone getting serious about patch management over at

In their explanation of the severity rating scheme, the Microsoft
Security Response Center (MSRC) said;

"One of our major concerns is that, all too often, customers fail to
install the security patches that would protect their systems"

Um, maybe we know why. Instead of providing a robust, accurate, and
incredibly reliable mechanism for getting patches onto systems,
Microsoft have provided a hodge-podge of services (WU), free versions of
fee-based tools with limited functionality (HFNetchk/MBSA), and myriad
"download" lists like;

Where each appear to have little knowledge of the others, often show
different results for the same system, or make it near impossible to
determine precisely what you need (particularly if you're not using all
of the latest versions of their software).

Then there are the issues like the difficulty correlating Microsoft
Security Bulletin patches to patches offered in other locations. Windows
Update now lists patches according to their KB number, but the Security
Bulletins today almost always say the KB will follow, or only show the
KB number on the download page. WU does embed the MSxx-xxx number in the
text, but its not obvious. Clearly they think they're talking to
different people and there's no common language.

I see the problem as two-fold;

1. There is no centralized organization, and therefore no comprehensive
and consistent plan, for patch management of Microsoft products. They
seem to be able to pick and choose between the various existing
services/download methods, or create an entirely new one of their own if
they want.

A good example is getting a cumulative roll-up patch for IE that doesn't
include patches for Windows Media Player. Or a cumulative roll-up for
IIS that doesn't include an Index Server patch. Why should they, they're
different products and under different managers with different
development teams! Unfortunately, most of us see them as linked and our
expectorations aren't in synch with what Microsoft is delivering.

There are, as far as I know, at least 3 different hotfix installation
tools out there. They use different switches to perform similar actions,
and none of them are well documented for the System Administrator. So
while its possible to extract the contents of a hotfix and examine the
individual files, using the /? Switch to see what you can do doesn't
show you the /x switch that expands the hotfix.

WU updates many products, but it doesn't update Office. HFNetchk checks
for missing patches in several products, but it doesn't know about
Exchange or Office. Office knows about its products, but it doesn't know
about IE, a crucial component of every Office product.

Is it really that difficult to bring this all together? The simple
answer is yes, it is. In my opinion, its largely due to the loose
management structure and lack of any focused emphasis on this as a
problem. Microsoft just don't see this as a big enough problem for its
customers to do all of the things required to fix it.

Numerous groups in Microsoft have attempted to fix this issue "once and
for all", with little success. Ultimately they created their own method,
causing further variation and confusion for System Administrators.

Trying to get consensus on this issue has failed, **for years**, its
time someone with sufficient clout (e.g. Gates himself) put their foot
down and mandated a plan of action that everyone *must* follow.

No doubt such a plan is in the works, but I also suspect that any such
plan is going to be based on .NET and only usable to customers who've
upgraded (or upgrade immediately upon release). So the solution will be
unusable by the *majority* of customers for years. One could easily
infer that this is a ploy to get people to upgrade..."if you want to be
secure, you must upgrade to .NET servers and Windows XP desktops,
otherwise your SOL.

2. Microsoft, and 3rd parties, have made numerous for-free programs
which "handle" this problem. We shouldn't be expecting this problem to
be solved for free. Free tools allow us to determine just how much of a
problem we have, but they can't solve the problem for us. For that, we
need to buy another tool, suite, or sub-system.

That's like saying that your car's warranty will get you a letter in the
mail when there's a recall, but it won't give you the repair. For the
repair you'll have to get the part and install it yourself, or pay
someone to do that for you.


As much as I hate saying that Microsoft should kill a product space,
this is one that *must* be killed. I want robust, accurate, and
incredibly reliable patch management for free, regardless of whether I'm
a home user or in charge of 70,000 systems. If for no other reason than
should Microsoft have to provide such a system, it will be an incentive
to minimize the number of patches they release (by minimizing the number
of patches their products need).

Another reason is because only Microsoft should be held responsible for
ensuring my system is patched. Microsoft make mistakes, as the problem
with FSCFG.DLL in MS02-018 demonstrated. Should a 3rd party be held
liable for such a mistake (your system doesn't get patched because the
patch is incorrect)? Who knows better than Microsoft whether a patch is
or isn't correctly applied to my system? Who can tell better whether I
have the affected software on my computer?

With "Trustworthy Computing" being such a hot topic at Microsoft these
days, they need to acknowledge that Sustained Engineering is one of the
most crucial elements.

The difference between what we have, and what we want, is the difference

- walking around and touching every machine in my environment versus
having confidence that a robust, accurate, and incredibly reliable
*free* patch management system has done it for me overnight.

- having my entire internal network shutdown because of rampant Nimda
attacks versus having the confidence they were all patched 5 months
earlier when the patch was released, and being able to verify they were.

- having my entire mail system shutdown because of a Badtrans.b
infection versus knowing that all of my IE installations were patched
within the 12 days we had after the patch release and before the first

The solution today is available only via money, or manpower, no Zero
Administration here (replace Zero with the number of dollars for a
solution or the number of people). The various free tidbits we're
currently offered only complicate the situation, because they're under
different management and provide inconsistent results. Fee-based 3rd
party solutions aren't the answer either, because they don't have all of
the information they need from all of the product groups who many
potentially issue their own patches (in their own format, using their
own detection methods, etc...)

So we await the next big attack, knowing full well that more than enough
machines will be susceptible to it regardless of when the patch for its
exploitation was made available. Not enough people have the money to
solve this problem, and most of those don't have enough people either.
Microsoft will say they made the patch available prior to the attack,
and try to duck the bullet again.

You would've thought that Code Red, and then Nimda, was enough to
convince them to go back and fix what's actually deployed, but alas,
nobody in Redmond runs Windows NT 4.0 any more so from their perspective
its not a problem. Few even run Windows 2000 any more. I expect that
robust, accurate, and incredibly reliable free patch management for the
systems that **will go down** as a result of the next big attack will
never come from Redmond.

Budget your upgrades to .NET and Windows XP now, or expect to be

...unless of course we can get Microsoft to listen. Unfortunately I
haven't had much success lately.

Russ - Surgeon General of TruSecure Corporation/NTBugtraq Editor