Posted on 02/17/2003 8:52:24 PM PST by Fractal Trader
Oracle admins are in for a busy time with the publication of no less than six vulnerabilities over the last week.
Four of the vulnerabilities are buffer overflow flaws affecting various components of Oracle9i Database Server. Then there's two flaws affecting Oracle9i Application Server, which pose denial of service risks... or worse.
Some are potentially very nasty indeed. Oracle describes them as critical and that's not the half of it...
The buffer overflows in Database server involve: the ORACLE.EXE binary, the TO_TIMESTAMP_TZ function, the TZ_OFFSET function and DIRECTORY parameter of Oracle9i Database Server.
These are explained in greater depth in the BugTraq advisories linked to above and the security section of Oracle's Web site.
The web site also gives more refers to two Oracle9i Application Server vulnerabilities (involving DAV_PUBLIC Directory and the mod_oradav Module)
All vulnerabilities were posted to BugTraq, and patched published by Oracle, last weekend. Over the weekend security researchers have been digesting these reports, and coming up with some potentially unsettling conclusions.
David Litchfield, of NGSSoftware, the security firm that has carved something of a niche for itself in unearthed Oracle flaws (and did the lion's share of the work this time too), tells us the majority of the Oracle9i Database Server require an attacker to have a valid user name and password.
So the greatest risk here comes from a buffer overflow glitch within the Database Server's authentication process, which a post from NGSSoftware to BugTraq today explains in much greater depth. Various flavours of Database Server (8i, 8.1.7, 8.0.6) as well as Oracle9i are potentially vulnerable to this attack, according to NGSSoftware.
Combine that with an Oracle9i Application Server Format String Vulnerability, and we have a way an attacker might gain control of Ap Server and get around what firewall rules might otherwise guard against attack against (potentially vulnerable) Database Servers.
Oracle describes this as only a denial of service risk but the issue, albeit it tricky to exploit, seems to go deeper than this would suggest.
Litchfield, in masterly understatement, says these various vulnerabilities "need attention".
Once again: Oracle's patches can be obtained via links on its Web site here. ®
As I sit here typing, I have a copy of the latest competition, Sun One, ready for install. Let's hope that MacNealy's crew hasn't erred in the same ways of Oracle and Microsoft.
Let me see if I can get this right...there are really two initiatives going on here, which I find hard to separate despite the fact that a colleague of mine is writing the technical specs for this. The broad characterization of the initiative is to create a unified platform to manage ALL IT infrastructure (including Linux and Windows and, of course, Solaris) from a single perspective by *virtualizing* the resources to a more abstract level.
The other way to look at this (I'm not sure if I have this totally correct) is that they are creating a unified "App Server" which will provide all basic internet services, but allow for the plug and play removal of certain functionality to be replaced by more advanced code. For example, it has basic certificate handling capabilities, but you can add RSA modules. It has a standard J2EE server, but you can add BEA or Web Sphere, and so on.
That's about the best I can do at untangling it. Sun has promising initiatives which they are undertaking, but has difficulties in getting out a coherent marketing messages. Dates back at least to "The Network is the Computer" slogans and the infamous, "We put the dot in .com" (or whatever silly thing it was)
Once upon a time, back before y2k, our Accounting group fell behind on patches for almost a year (the DBA was subsequently fired).
Between the DB and apps, there was almost a hundred patches they needed to apply.
This is nothing new, unfortunately.
'Big Software' houses have, at best, shady business practices. The general rule of thumb -- never use a brand-new product from *anyone*.
Or MS, right?
Never use a first release of anything for real work. Using 'bleeding edge' technology will always be more costly, and fraught with bugs and issues. It's the nature of the beast.
We're still on Oracle 8 at work.
As I've said over and over again -- I wouldn't have suggested using Java for any 'real' work before about 1999, after it had been in general use for about 5 years.
Building 'web services' in Java is not a new thing, fortunately. The technology is very mature, and ready for prime-time.
.NET, on the other hand . . .
I've been building 'web services' for years, in Java.
The new toolkit is just that, a new toolkit. A new way of doing it, not the only way.
You'd really think I'd have learned about talking tech with people under 18 . . .
Yeah, sure, right.
You're not going to claim the MS invented 'web services' now, with .NET, are you?
A Web service is a software system identified by a URI [RFC 2396], whose public interfaces and bindings are defined and described using XML. Its definition can be discovered by other software systems. These systems may then interact with the Web service in a manner prescribed by its definition, using XML based messages conveyed by Internet protocols.
Been building them for years. Servlets, in Java. MS is, once again, the 'Johnny Come Lately' here.
As you know, all my tools are deployed *internally*, behind the firewall. And the industry clearly hasn't decided that they're secure enough to deploy on the internet. They aren't exactly dominating web development.
Are you going to promise yet again that MS's new offerings will be 'secure'? Let's see, how many times does that make now?
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.