Oracle Too Expensive At Any Price

“Oracle’s new cloud licensing policy [PDF] says an AWS vCPU is now treated as a full core if hyperthreading is not enabled. A user renting two AWS vCPUS therefore needs to pay full freight for both, effectively doubling the number of Oracle licences required to run Big Red inside AWS. And therefore doubling the cost as well.”
 
See Oracle effectively doubles licence fees to run its stuff in AWS
If you can’t afford to convert to PostgreSQL for the extra amount Oracle will charge you, you’re not doing IT properly. Change. If you can afford to pay twice for stuff, just retire. You don’t need to work at all and can coast on your wealth.

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 and tagged , , , , , . Bookmark the permalink.

7 Responses to Oracle Too Expensive At Any Price

  1. oiaohm says:

    http://mrpogson.com/2017/01/26/arm-and-intel-go-head-to-head-in-uk/#comment-369509
    DrLoser so like here were he did not say a thing because the reality the answers right. Wait he did not post a single comment about it. Nothing surprising there.

  2. DrLoser says:

    Delight! You are back again, Fifi! How very sweet of you to grace our presence!
    I particularly look forward to Dougie destroying every single stupid “point” you claim to make.

  3. oiaohm says:

    kurkosdr “Enterprise Legacy System”

    That is the thing.
    https://www.enterprisedb.com/database-compatibility-technology-oracle
    Its cheaper to pay for enterprisedb wrapper than what it is for orcale for the same performance.

    As I said, many Enterprise Legacy Systems are built with certain assumptions about how the query optimizer works, what query optimizer directives are present and how they work, what use-this-index directives are present and how they work, and Zeus knows what else.
    That is exactly what the enterprisedb emulation does with paid subscription to deal with what ever issues crop up. Yes updating version or Oracle from a insecure version to current secure version can also cause legacy systems to have issues.

    So there is a big question why in heck are people paying the Oracle license fee when enterprisedb license fee is cheaper todo the same thing and gets your data stored in postgresql so you can migrate forever over time away from Oracle.

    It costs dollars at times to dig your way out of particular made holes. From Oracle to postgresql there is a well taken path of middle step of enterprisedb.

    Hint: support
    Is in fact a reason to go enterprisedb or an postgresql. There was a formal compare between speed of issue fixing between people paying for postgresql support and those paying for orcale. The figure is scary. 1/20 the time to get issue fixed with postgresql based solutions.

    http://www.oracle.com/us/corporate/contracts/license-service-agreement/index.html

    There is something critical missing where is the SLA decribing how much it costs to have oracle developers respond inside a particular time frame no matter what. You can get this out of postgresql paid support you cannot get this for Oracle db. Yes paying way less and get that your issue by contract has to be addressed in a timely manner. So its Oracle db that is without quality support. Of course postgresql quality support is not free.

    You might find something that looked like it until you wake up its Oracle who defines the severity of the issue not you the customer and they nicely in their documentation don’t mention severity level 4 what is basically Oracle language for won’t fix. This is something different about Postgresql where paid support where customer define the severity level based on how much the issue is effecting their operations. Yes Oracle also reserves the right to change the severity level of a bug report so you report something severity level 1 they can move it all the way to severity 4 and never breach contract. Nothing like paying for support that is only smoke and mirrors. So the Oracle SLA is your issue will be fixed maybe quickly or maybe never. So not much better than using postgresql without paying for support.

  4. kurkosdr says:

    (not that Pog understood anything I said of course, and if he did he probably thinks companies are stooopid to use those non-standard features. Pog is the (d)evolution of your stereotypical 2000s freetard who recommends productX as the one-size fits all for needs despite only running a basic webblog and warez site himself)

  5. kurkosdr says:

    Hint: support.

    That’s another valid point, but even if the FOSS databases offer you an excellent support option for $1 buck a month and employ telepathy to diagnose your problems the moment they happen, it is still not enough.

    As I said, many Enterprise Legacy Systems are built with certain assumptions about how the query optimizer works, what query optimizer directives are present and how they work, what use-this-index directives are present and how they work, and Zeus knows what else.

    If you are hosting some basic blog and use plain SQL then you can effortlessly move from db management system to db management system, but it is madness to suggest to an enterprise whose Enterprise Legacy System is locked-in to Oracle DB or Informix to do such.

    This is also why companies like Oracle and IBM support standards even though they have a financial interest in their products having non-standard features which are exclusive. They know very well that the features the standards cover are inadequate (and the bureaucracy surrounding standards committees make sure they will stay that way for any given point in time), so they know customers will have to use the non-standard exclusive features to do business and that they will get locked in anyways. It’s the same thing with OpenGL, with iOS and Android having their own extensions to get business done. Or the endless extensions to HTML, CSS and JS (such as Google’s FileWriter or the inconsistencies in the BOM).

  6. Deaf Spy says:

    Doing IT properly is something, Robert, you know of as much as of Windows 10, i.e. basically nothing.

    Hint: support.

  7. kurkosdr says:

    If you can’t afford to convert to PostgreSQL for the extra amount Oracle will charge you, you’re not doing IT properly.

    Of course, Pog doesn’t realise that many companies have something known as the “Enterprise Legacy System” which makes some important part of the business work. The main characteristics of an Enterprise Legacy System is that nobody understands the whole of it (but some people understand parts of it), that it has gathered many important business rules, edge cases and performance optimisations over the years that aren’t 100% documented and hence the system would be a pain to rewrite, and most importantly, it is built around and optimized for some database management system, and hence the code makes certain assumptions about certain quirks and exclusive features of the database management system being present. It also makes assumptions about the performance of the database management system and that certain query optimization directives are present and well-supported.

    I had the misfortune of coming across such a system at work (just a glance though). It takes care of some important business logic for 6 customers of the company. The product consists of a “core” module and 6 “adapter” modules that add or modify to the core business logic in various ways (calling a core method and changing the result, replacing a core method, replacing the queries of the core method using a shared object or by passing the query as an argument). The whole thing is built around IBM Informix database and is built with the assumption that certain query optimization directives and use-this-index directives of Informix are well-supported and functioning exactly as expected. Just the logs it outputs at INFO level when performing the functional tests are 47 MBs long.

    To give you a sense of the massiveness of the tool and the risk involved, I know a guy (who is supposed to know the system) who is tasked with finding out why the Enterprise Legacy System outputs at FULL DEBUG level (despite being told to output at INFO level) during the tests when built in Jenkins (but not when built locally), which causes gigabytes of logs to be created and causes the Jenkins job that builds and tests it to ran out of workspace space and crash after hours of running the build job.

    Rewriting such a tool would be a massive undertaking of high risk and cost, so it is more prudent for the company to just pay IBM what IBM wants to use Informix than move to MySQL and PostgreSQL or some other FOSS database. Same for Oracle customers paying Oracle what it wants to use the Oracle database. But I am sure our company’s CTO could use your invaluable advice Pog and make strategic decisions around it, because you are clearly in position to make blanket statements like “you’re not doing IT properly. Change. If you can afford to pay twice for stuff, just retire. You don’t need to work at all and can coast on your wealth.”

    PS: Of course, new tools are not built like this, it’s all serverless in AWS, but the license for Informix for the Enterprise Legacy System still needs to get paid.

    PPS: Oracle. We used to have customers, now we only have hostages ™

Leave a Reply

Your email address will not be published. Required fields are marked *