Re: Call for Oracle Licensing reform

Date: Thu, 8 Jan 2015 12:07:25 -0500
Message-ID: <>

It sounds like your database rep has no clue what he is talking about. Be wary of ANY statements made by your sales rep -- under the "Complete Agreement" section in the OLSA, such statements are completely non-binding and offer you no legal protection whatsoever in a future dispute (or audit().

There are no (true) objective rules for what does or does not constitute "hard" partitioning. Oracle decides this almost on a case-by-case basis, and for any purpose I am aware of, there is no customer-accessible repository of these decisions.

One option you (of course) have is to REMOVE a processor from you machine. Then the partitioning question vanishes, and your licensing situation is cut-and-dried.

Partitioning technologies like Xen (or Oracle VM) can be either "soft" or "hard", depending on (exactly) how you use them. An installation that is legally licensed today can become non-compliant tomorrow, if a sysadmin changes the configuration in the hypervisor without your knowledge.

Standard Edition licenses are relatively cheap, so you don't have huge stakes involved here. It sounds, though, like the hardware in question is inexpensive, also. You may want to decide whether partitioning is worth the trouble (and risk) before you even start. If you are sure that it is, you might perhaps want to involve LMS from the very start, and get a written statement from them that your deployment is compliant. Ideally, you would want a formal ammendment to the OLSA stating this, but your chances of getting such a thing for a single socket of Standard Edition are probably about NIL.

On Thu, Jan 8, 2015 at 11:52 AM, Ronan Merrick <> wrote:

> Hi Listers,
> I bought a new server with two cpus. I want to partition this machine and
> just use one cpu for licensing purposes. I was reading around the
> documentation and got immersed in the virtualization debate.
> My understanding is that hard partitioning is only recognised for
> licensing purposes. As far as I can see the nitty gritty of hard
> partitioning is that you can partition at the core level which would seem
> to be relevant for the enterprise edition.
> The database I want to deploy on the server is standard edition so
> shouldn't soft partitioning apply since this is on a socket basis not the
> number of cores?
> When I bought the server the Oracle hardware rep told me that the xen
> hypervisor on Oracle VM would be adequate to ensure license compliance if I
> wanted to partition the machine.
> However my Oracle database rep he indicated that I may have to license
> both cpus as partitioning is not recognised for standard edition. He didn't
> sound 100% sure though and the hardware rep at the time said that the
> database guys don't like them making people aware of the partitioning issue.
> Just wondering if anybody has any experience of this
> Ronan
> On 8 Jan 2015 06:34, "Paresh Yadav" <> wrote:
>> It is interesting that you mention Oracle SE and I agree that there are
>> times when SE is all one needs. Cost of migrating away from Oracle might be
>> high but cost of starting with alternatives is not that high if one knows
>> what they need.
>> You are missing the NoSQL bandwagon as there are many applications where
>> ACID compliance and other bells and whistles available with Oracle database
>> is simply not needed.
>> Thanks
>> Paresh
>> 416-688-1003
>> On Thu, Jan 8, 2015 at 1:25 AM, xxxxxxxx wrote:
>>> There is no need for confrontation. I have simply made a statement of
>>> fact. "Audit" and "License" are not synonyms.
>>> As for how long I have been involved in licensing Oracle's products?
>>> Well, I have never been an employee of Oracle, nor have I actually
>>> *written* one of their license agreements. I have, however, been
>>> *using* Oracle databases for close to 20 years, and reading the license
>>> agreements and advising customers on how to (correctly) comply with their
>>> licenses for much of that time. I am not a lawyer, but I am fairly widely
>>> perceived as "quite knowledgeable" on this subject. (Sadly, though, much
>>> of that perception may be due to the fact that few others actually
>>> *read* Oracle's license agreements, as I do.)
>>> In any case, modifying "audit" policy is not in any way equivalent to
>>> modifying a license agreement. Oracle could cease auditing entirely, and
>>> it would have NO AFFECT on licensing. (It may, however, have an affect on
>>> Oracle's revenue.) A license is a binding contract between two parties.
>>> "Audit" is simply one of several means of ensuring that one or the other
>>> party complies with that contract. Changing the way you audit does NOT
>>> change the contract. Changing the CONTRACT is the way you change the
>>> contract.
>>> Its funny that you mention "alternatives" in this discussion. There are
>>> -- of course -- many. My favorite by far, happens to be Oracle. Oracle
>>> Standard Edition, that is.
>>> Under the current licensing terms, Oracle Standard Edition (One) can
>>> cost 1/50th as much as Oracle Enterprise Edition, or less, if you happen to
>>> use Named User licenses. Every time processor vendors increase core
>>> counts, this cost factor increases. In just a few years, I expect the
>>> *minimum* cost for an Oracle EE database server will be 100x to 250x
>>> the cost of an Oracle SE database server.
>>> When you consider the cost of migrating applications from one database
>>> to another, the very modest price of Oracle SE-One can compete very
>>> effectively even with "free" products like MySQL. As a rule, the effort
>>> for migrating applications from Oracle to MySQL is usually far from "free".
>>> On Wed, Jan 7, 2015 at 10:33 PM, Paresh Yadav <> wrote:
>>>> xxxxxx,
>>>> I don't like to take a confrontational stance so pardon my position
>>>> here. How long have you been involved with licensing Oracle products?
>>>> On the audit side, whenever you play a game ( let us say Monopoly as
>>>> most of us should have played it at some point time or a close clone in
>>>> disguise) there are rules and there are strategies. Audit is the strategy
>>>> to make licensing rules work. Nothing wrong or illegal on Oracle's part but
>>>> I hope they realize that brute force licensing is hearting Oracle's long
>>>> term prospects ( who in corporate world cares for longterm prospects when
>>>> exec bonuses and stock price depends on next quarter numbers). I am sure
>>>> Oracle has gurus doing number crunching and may be using predictive
>>>> analysis to determine the equilibrium point or point of diminishing returns
>>>> as they say in economics and trying to stay in optimal zone when it comes
>>>> to squeezing all possible revenues. My empirical observations in last 5
>>>> years say not only startups but even fortune 100 orgs flush with cash have
>>>> shunned Oracle over other alternatives.
>>>> I love Oracle database tech and consider it one of the best creations
>>>> in software enginnering even though bugs galore. But now there are
>>>> alternatives that can meet the need at fraction of the Oracle database
>>>> licensing costs and even the 'no body gets fired for buying Oracle' crowd
>>>> is using alternatives in the prod as IT becomes commoditized and is being
>>>> looked as a cost centre more than competitive advantage.
>>>> Amen!
>>>> Paresh
>>>> On Wednesday, January 7, 2015, xxxxxxxx wrote:
>>>>> No. The audit process is how the existing licensing system is
>>>>> *enforced*.
>>>>> Reforming the audit process has no actual effect on the terms and
>>>>> conditions of the license agreement -- it would affect only how and when
>>>>> that agreement is enforced and how people feel about that enforcement after
>>>>> the fact. The article provided suggest that there is a great deal of room
>>>>> for improvement in these things -- especially "how people feel", but no
>>>>> amount of change or improvement to the audit process in any way changes the
>>>>> license. Even if Oracle announced that they would NEVER audit any customer
>>>>> EVER again (which would certainly avoid lots of hurt feelings) honest
>>>>> customers who choose to honour the license agreement (and this is probably
>>>>> the vast majority) would still have the same terms to comply with.
>>>>> The last time Oracle made any sort of major reforms to the license
>>>>> agreement, they removed outdated metrics like "Power Unit" (number of CPUs
>>>>> times clock rate) that forced exponential cost increases on customers
>>>>> (because clock rates were doubling every 18 months, even though actual
>>>>> performance was not keeping pace) and other metrics like "Concurrent
>>>>> Users" which were next to impossible to measure and validate. (That is,
>>>>> LMS probably found "concurrent users" hard to count, while customers found
>>>>> it almost equally difficult to demonstrate their compliance.)
>>>>> On Wed, Jan 7, 2015 at 2:05 PM, Paresh Yadav <>
>>>>> wrote:
>>>>>> Audit process is actually licensing process in disguise.
>>>>>> Those who day dream about getting it reformed -
>>>>>> :).
>>>>>> Thanks
>>>>>> Paresh
>>>>>> 416-688-1003
>>>>>> On Wed, Jan 7, 2015 at 1:56 PM, MARK BRINSMEAD <
>>>>>>> wrote:
>>>>>>> Indeed.
>>>>>>> Actually, the article seems to be mainly calling for Oracle to
>>>>>>> reform its *audit* practices.
>>>>>>> Personally, I have been curious -- for a very long time now -- as to
>>>>>>> how long it will take Oracle to revise its license metrics. Now that
>>>>>>> processor designers are scaling performance by increasing core counts
>>>>>>> rather than clock rates, the cost Oracle database servers is growing at
>>>>>>> (what is likely to be) an exponential rate. Likewise, the price gulf
>>>>>>> between EE and SE products is also growing at the same sort of rate.
>>>>>>> On Wed, Jan 7, 2015 at 1:49 PM, Hans Forbrich <
>>>>>>>> wrote:
>>>>>>>> While interesting .... "be careful what you wish for"
>>>>>>>> /Hans
>>>>>>>> On 07/01/2015 11:39 AM, Dennis Williams wrote:
>>>>>>>> List,
>>>>>>>> This is an interesting article on how Oracle needs to reform its
>>>>>>>> licensing practices:
>>>>>>>> Dennis Williams
>>>> --
>>>> Thanks
>>>> Paresh Yadav
>>>> 416-688-1003

Received on Thu Jan 08 2015 - 18:07:25 CET

Original text of this message