"The only apparent requirement is that the customer follows through and commits to using Oracle’s software instead of that of rivals."
Er, didn't Microsoft get spanked (eventually) for doing more or less this?
Oracle customers could soon be the beneficiaries of an unlimited, all-you-can-eat licence for its core database, with the giant understood to be readying a deal that would grant use of its database in perpetuity at a flat rate. Oracle’s prospective licence is being referred to as the Perpetual User License Agreement (PULA). …
Microsoft got batted because they were the dominant player in one market and were trying to use that muscle to move into another market. Now if the new license is loss making, and oracle have a dominant position in the database market (Over a certain market share) then they could be in trouble however there are plenty of other players in the commercial database market including microsoft and IBM.
I take it you don't understand large scale databases very well in general then.
I won't bother dissing mysql, which is fine for 90% of its use cases, but seems to fall down on really complex stuff and has questionable ACID compliance.
But let's take postgresql, which I love. It seems as if the databases can grow to 60-70 gb without any hassles. No, not talking about theoretical limits, more what runs comfortably in real life. But, what if you need to grow 10x above that? Then you may want to go back and look at Oracle or DB2. Or even, who knows, with much caution, look at MSSQL (which isn't super cheap when you get into the real enterprisey versions and sizing is also an issue). And what if your OS platform or 3rd party application isn't supported by postgres? Again, you may need to shell out the big $$$. Then there is the availability of really skilled DBAs. Many know Oracle very well. Postgres? Maybe not as many.
It all depends on your needs. I would definitely go for postgres at first (or sqlite under special conditions). Or even mysql if I wasn't doing anything fancy. But these may not always fit the bill and it may end up being an 80/20 problem where you solve almost everything but fall down on the remainder. Oracle costs big $ partially because they are one, if not the, top db vendors.
Point is, sometimes it is cheaper to buy something expensive rather than paying devs to code workarounds on the wrong tool.
As to this "all-you-can-eat" offering ending up being a good $ deal for customers? I'll believe that when I see it.
> the user will pay a fee – calculated by Oracle estimated on the size of your estate
> It would ... eliminate the potential for nasty surprise
Yeah right. Instead of *counting* how many licences you actually use, they get to decide how many they *think* you might be using. Or: every year you have to negotiate afresh to come up with a figure agreed by both sides.
Sounds like a way to increase your exposure to risk, rather than reduce it.
Or maybe it's like this:
- you pay a fee based on your current counted usage
- you are safe and secure in the knowledge that during the year you can deploy more instances without having to buy individual licences
- instead, you are stung at the end of the year when it's time to count the requirements for the following year
If so, it just means you get (up to) one year's free trial deployment, after which you pay the usual rates. Standard marketing trick.
The issue with Oracle was not seat/processor licenses only, often it was "options". Oracle comes with several built-in features you can't use until you get a formal license for that "option". Exactly, features are already there, you don't need to install an add-on or enter a new license key, just you can't use it without a piece of paper stating you can do it. Crazy? Yes, but that's how it works.
One is (was?) for example partitioning. If your DBA doesn't know Oracle licensing well enough, and starts to think partitioning some large tables is good, and does it without an option license, an Oracle audit will spot it and ask you money. Other options include(d) Advanced Security (encryption of data and communication channel), RAC (clustering), standy databases, some tuning/diagnostic features, and so on. Just a few are true add-ons to install separately (my last big experience with Oracle licensing was with 11g, thereby something may have changed since 12c)
Thereby having a flat fee letting you use any option if you start to need it without going through another round of licensing talks, would be welcome by many DBAs.
"An unlimited, all-you-can-use, flat licence based on an estimate would be a major departure for Oracle. It seems it's actually willing to give up on accurate measures of what customers are consuming and to trade that for an estimate."
The question that interests most customers will be, Do I pay more, or less?
Here's a completely radical idea for Oracle... How about they make 1 product that works than 4 competing products that don't. I deal with Salesforce & Dynamics on a daily basis...in the last few months I've been subjected to Oracle On Demand (which is crap), Oracle Fusion (which is diabolical..it shouldn't take me 2 days to work out how to edit some fields and links), I'll be working on Siebel shortly (and expect that to be shit as well) and today I discover they also do something called Sales Cloud...
What possible need is there for 4 systems that all do the same thing but each with differing degrees of rubbishness.
Oracle Fusion is just the merge that you are asking for. The reason the older application products are still supported is because customers demand not only long support guarantees, but continued upgrades even for the previous generations of application software.
Oracle's acquistions are often for the better technology, and those improved technologies are pretty quickly merged - e.g. Oracle Discoverer was superceded by the BI component acquired with Siebel, Oracle Warehouse Builder was superceded by the purchase of Sunopsis, Oracle App server was replaced by BEA weblogic, even the database has many internal components that were acquired (e.g. data mining, OLAP). These merges don't happen overnight, but are usually executed gradually over 2-3 major versions with co-existence to limit the impact on customers who are still using the older product.
I'm afraid your showing your ignorance here, which somewhat undermines the validity of your criticism.
'On Demand' is Oracle supplying SaaS for any application, so that's a customer deployment choice.
Oracle Fusion is the name for the suite of the new generation of application products which supercedes the various acquisitions (e.g. JD Edwards, Peoplesoft, Siebel)- and Sales Cloud is one component of that. However, Oracle continues to support and update those older products - see apps unlimited.
So the four systems are not competing systems and even for the two systems that offer similar capability, that's to support customers on older versions.
If you were to think a while about your statement, you would realise that having more than one version of major applications is inevitable for any mature company that has been delivering applications for more than one major software language shift and continues to provide support for the customers lifetime use of the product. Without this approach, all our apps would still be written in COBOL.
Salesforce only has one version because it hasn't been around long enough to experience one of those shifts yet.
If this only covers the database, I wonder what they're going to do for all their other messy enterprisey applications (Financials, PeopleSoft, JD Edwards, Siebel, WebLogic, Solaris, and so on).
Also, the other thing that isn't mentioned is the support entitlements. I set up Oracle RAC for a proof of concept in our lab a couple years ago. I don't know how i could have completed it without falling back on our company's paid-up Oracle maintenance contract. The released software (available for "free" on their site) was not usable (read: would not even install properly) without several patches, only available from My Oracle Support. The documentation had errors in it, which were highlighted in several linked "support notes" and also only available on MOS.
Seriously, the entire Oracle infrastructure is designed around the expensive consultant model. I wouldn't be surprised if they raise the support rates to compensate for it.
You should get an Oracle Partner subscription (hope it still exists). It allows you to use the licensed products for development/test/demo without issues, any number of instances you need, and you have access to support for patches and service requests. What you can't do is use that license for any production installation.
The subscription is a yearly fee, it's very much alike MSDN and the like, designed to keep dev/test labs to a manageable budget.
It's not perpetual all you can eat. You have to pay for each user each month - it's a dream come true for Oracle, and one they're trying to move to by selling Oracle Cloud Services.
Salesforce runs only on Salesforce servers giving them total control over whatever part of your business you move onto Salesforce. There's no need for them to audit your software use - they can monitor it minute by minute and shut you off if you don't pay up.
This sounds a lot like the Microsoft Campus agreement in academia, where they charge a fee based on the number of computers on site (regardless of what software they run). This has an obvious advantage: all your systems are licensed for the relevant software, without any further checks. The downside, of course, is that every system is then licensed for MS software - even if it's actually being used for Linux, Solaris, *BSD or as an embedded control system for something - at which point, there is no saving in using a rival product: you've already paid for the MS option anyway.
I remember coming across this with Oracle in another university years ago - asked why they were using Oracle for an internal user database (something quite trivial, a few Gb at most), the answer was that they were already licensed for it, so why not? Of course, next time the Oracle license is up for renewal, that's one more critical system relying on keeping the license current...
Even with "perpetual all-you-can-eat Database" subscriptions, Oracle solutions will still be an order of magnitude more costly and convoluted that modern, powerful and equally scalable solutions based on products like Postgresql object/relational RDMS, and with world class expertise and support from solution providers such as RedHat, IBM and EnterpriseDB, Fujitsu and many more.
Many expert technologists and business visionaries were correct in their assessment that it was a severe misfortune for all great technology when Oracle purchased Sun Microsystem and the plethora of fantastic Free/Open Source Software (FOSS) software included in the transaction, that Oracle is attempting miserably to make fully proprietary - like Java, MySQL, VirtualBox and other aplications.
Crass greed combined with draconian proprietary software - even most that are inferior to better FOSS, are badly hurting all technology development, particularly here in USA.
Postgres should really start to support unstructured data (LOBs) far better if it wants to replace anything. Real scalability, is another issue.
About MySQL it's funny how many people never understood it was "free" only for GPL applications. It was always a commercial database for non-GPL applications.
Also Oracle purchased Sun because Sun went nowhere - but bankrupt - with all that marvelous FOSS software - why? Because sustaining a business when everybody just wants to use your software for free doesn't take you very far, if you have to sustain the full costs of developing it?
"severe misfortune for all great technology"
"Crass greed combined with draconian proprietary software"
MySQL was sold to SUN. That was misfourtune. That was greed by the original founders.
Then the very same person goes ahead and launches MARIADB.
Everyone hails him. Thats hypocrisy.
Lets begin by saying something to Monty first.
Java Solaris and many softwares were born in Sun.
Sun just lost the focus after Khosla was kicked out.
Scott McNealy was unable to look 'Beyond 2000'.
Regarding Oracle, they have a vision.
Vision thats based on revenue generation.
MS IBM HP SAP ORACLE all are 800 poound gorillas. they understand this and survive to this day.
In time they will adapt to the Cloud as well as IOT.
Johnathan Schwartz is a douch3b4g. Blogs and hip haristyle dont generate business.
Laser like focus does.
The ones that lack CEO quality talk hip and destroy companies.
The ones that do, buy a yatch.
To: anonymous Coward - responding to my remarks on PostgreSQL as a viable replacement for expensive Oracle database solutions, please refer to this article from EnterpriseDB (who employ core PostgreSQL developers) at:
There are also other technical documents indicating support for 'unstructured data' - similarly to functionality in most No-DQL databases, in PostgreSQL since release version 9.2.
MySQL GPL version is Free - as in Freedom, even when used with 'proprietary' applications. The commercial (paid for) release supports Oracle transaction engine, and offers 'commercial' technical support services, as well as certain proprietary utility functionality and "up to date" bug fixes that Oracle provie for FOSS version at much later date.
Finally, Oracle purchased Sun Microsystems 'primarily' to gain control of Java from strong influences of IBM and other competitors in the Java ecosystem. Sun's FOSS software policies which did have "some" negative effect on company's finances - which were mostly poor business practices, but had very little or nothing to do with main reason for Oracle purchase.
Please get "factual" information on technology issues 'before' posting - and especially anonmously.
After sitting down with their licensing people in 2014 and then having to pull out CPU's fro our Blades, we went through an audit in December and they said we owed them and exorbitant amount, so we sent them a nice long document telling them, "you told us this, so f*ck off" and 6 months later we are still awaiting their reply.
So let me get this straight, you have unlimited deployment rights but it is based on an assessment of the size of your estate. You pay every year based on this, and I assume you would have to reassess each year.
So forgive for probably sounding stupid but this sounds more like a perpetual subscription than anything else and also suggests that you are effectively going to have to audit yourself every year to assess your estate size.... which after a year or two of unlimited deployment, don't give a crap about clusters or vSphere 6 type apathy because hey you have unlimited deployment, your estate is suddenly bloated beyond all believable bounds.
Biting the hand that feeds IT © 1998–2019