Oracle Sinks

Oracle is the goto guy of enterprise databases. Data is the anchor for corporations with global reach and large organizations have clusters of databases for security, performance and to deal with the complexity and scope of operations. That dependable Oracle database has been holed below the waterline.

Infoworld has discovered that there is a fundamental flaw with the counters that allow transactions on Oracle database to be synchronized around the globe. There is a Y2K-like error in the way backups and some transactions cause the counter to be incremented towards the upper limit of the counter’s value. The result is an intruder using routine commands could break the database. The larger and more nodes the database has the bigger the vulnerability. Basically a storm of increments to allow for synchronization can rather quickly bring the counter to its limit.

There are a couple of fixes for this problem: a temporary shutdown to reset the counter and/or a patch that Oracle has developed. Either are a costly interruption in the service upon which Oracle has built a saleable reputation.

This is another example of how IT with a monoculture of software can leave itself open to serious threats. That even the normal backup procedure relied upon for the ultimate security layer is a part of the problem must be giving system admins nightmares. I would bet there are a lot of Post-itTM notes up today logging the idea of rethinking the databasery of large organizations and aspiring smaller organizations. I would bet some are considering PostgreSQL or dual-database systems to close out the possibility of database-Armageddon in the future. I would bet a few intruders will find unpatched systems out there with which to create some chaos.

The ultimate blow to Oracle’s reputation in all this is that Oracle was aware of the problem and assumed users would never find it. They were counting on security through obscurity.
“After much discussion and exchange of technical data, Oracle acknowledged that there were ways to increase the SCN at will. Referring to one method, Townsend said, “This is an undocumented, hidden parameter, so it was never intended for customers to discover and use this.”

However, we pointed out that there were several other methods that could be used; we sent those to Oracle as well.”

Is that good enough for a licence that costs £31,839.00 / Processor?

Once again, we see that dependence/lock-in to a single source of supply for anything in IT can be fatal. We saw that in Wintel (both costs and malware), hard drives made in Malaysia (flooding interrupted supply), and now databases. IT systems need to be robust and flexible which is not what lock-in gives.

For those considering PostgreSQL, you might be interested in the offerings by EnterpriseDB.
“EnterpriseDB is the only world wide provider of enterprise-class products and services based on PostgreSQL, the world’s most advanced and independent open source database.

Postgres Plus Advanced Server provides the most popular enterprise class features found in the leading proprietary products but at a dramatically lower total cost of ownership across transaction intensive as well as read intensive applications. Advanced Server also enables seamless migrations from Oracle® that save up to 90% of the cost of typical migrations.”

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology. Bookmark the permalink.

13 Responses to Oracle Sinks

  1. kozmcrae says:

    “Linus and his staff can “see” what is going on with Linux and could be effective in reviewing new code. Ditto for the various team leaders and developers at Microsoft”

    You’re a riot Clarence.

  2. Clarence Moon says:

    “It does not matter how many eyes there are if they cannot see”

    It is important to define the meaning of “see” I believe. It is not enough to be able to view. The viewer must have an undertanding of what is being viewed. That only comes with substantial design and development involvement of the software product itself. Linus and his staff can “see” what is going on with Linux and could be effective in reviewing new code. Ditto for the various team leaders and developers at Microsoft

  3. Kozmcrae says:

    “M$ receives millions of bug reports and they disappear into a black hole.”

    A bug report is not going to do Microsoft much good if the fix will create more problems. I would guess they concentrate their efforts on known security issues. Let the customers suffer with nonessential bugs.

  4. Clarence Moon made me laugh, writing “there are not any more useful eyes on the problem with FLOSS due to open source as there might be on a closed source project “

    That’s provably false. It’s mostly employees of M$ who get to peek at that other OS while millions get to peek at GNU/Linux. It does not matter how many eyes there are if they cannot see. With FLOSS we know they can see. It’s in the licence…

    While it is true that 90+% of FLOSS users use binary packages and do not look at the source code we do know there are many millions of FLOSS users who do look at the source code. We can even communicate with them because bug report lists are social sites. M$ receives millions of bug reports and they disappear into a black hole.

  5. Kozmcrae says:

    “Because 9.9 times out of 10…”

    That still beats the Microsoft coders who have no idea what kind of horrific changes their “fixes” will have on the Windows spaghetti code. Besides, you just made that number up. I fits your reality. The reality of the Cult of Microsoft.

  6. Clarence Moon says:

    It is obvious that the FLOSS advocates here do not really understand the way that major commercial software products are produced. The sense here is that there are many people reviewing the source code an thus the quality is improved and defects are found early with FLOSS. At our place code is intensely reviewed only near release dates to ensure that fixes to one problem do not cause a defect in another area.

    This can only be done by programmers who are expert in the area being fixed as well as other areas that may be affected. The casual FLOSS user is not at all capable of doing that and so there are not any more useful eyes on the problem with FLOSS due to open source as there might be on a closed source project wherein those working for the product company do have access to the source.

    All that a FLOSS user can effectively do is report that some release has a bug. They are no different than any end user in this regard. The truth about FLOSS is that there are far few “eyes” looking at, say, Open Office than there might be looking at Microsoft Office.

  7. oldman says:

    “What is generating a bug and what does your statement have to do with many eyeballs discovering errors in code?”

    Because 9.9 times out of 10 those discoverers have zero understanding about how their fix will effect other code. regression testing for the most part is almost non existent in the commune.

  8. Kozmcrae says:

    “Other than as a random bug generator…No.”

    What the hell are you saying? What is generating a bug and what does your statement have to do with many eyeballs discovering errors in code?

  9. oldman says:

    “What kind of proof are you looking for Dr. Loser? You don’t believe this example gives ample proof of the advantage of many eyeballs?”

    Other than as a random bug generator…No.

  10. kozmcrae says:

    “Or did you seriously believe that the massed ranks of FLOSS…”

    I’m not sure what you are talking about. The customers found the flaw that Oracle thought never would be found. This isn’t a FLOSS issue but it does demonstrate that given enough people, flaws will be found that otherwise wouldn’t. At least Oracle thought it wouldn’t be found. They underestimated the power of many eyeballs.

    What kind of proof are you looking for Dr. Loser? You don’t believe this example gives ample proof of the advantage of many eyeballs?

  11. Dr Loser says:

    @Koz:

    Well, Winter is proceeding in its usual way, and I’m feeling chilly. Maybe I’ll toss a few more eyeballs on the fire. God knows, Linux has conclusively proven that they are good for nothing else.

    Or did you seriously believe that the massed ranks of FLOSS, in direct disobedience to Robert (who has gone on record as never wanting to use Oracle stuff again), slaved over this proprietary system and figured out its faults?

    Ya gotta stay consistent if ya want to stay qualified as a Loon.

    @Robert:

    I like the Y2K analogy. Didn’t mean a thing, didn’t cause much more than a ripple, and nobody had any reason to care in the first place … chuckle.

    Or, alternatively, you could support the right of every snake-oil salesman out there to charge ridiculous amounts of money for non-solutions to a non-problem.

    The more I think about it, the more it looks like a 100% perfect analogy to me.

  12. Yes. That’s a shocker. I am terrified by security concerns just running a few machines on a LAN. Some of these enterprises have multi-gB databases holding the Crown Jewels. To have it shut down/inaccessible even for a few hours must be disastrous.

  13. Kozmcrae says:

    “…so it was never intended for customers to discover and use this.”

    Never underestimate the power of many eyeballs.

Leave a Reply