Jul 4: AWR Licensing
I love AWR, ASH and I even have an increasing, slightly grudging and cynical respect for ADDM. In fact, I'm in the course of putting together a two-hour training/technical mind-share session on their use at my current site because I have the joy of using them! Nothing would make me happier (in the Oracle technical arena) than for Oracle to include these features in the standard Enterprise licence so I support the letter and hats off to Mark Brinsmead for adding to Niall's (and others) earlier work. But it would be boring for me to just join the band-wagon and repeat what's already been said, so let me focus on some minor areas of difference.
1) The implication in Mark's blog is that adoption levels are almost zero, but that's not my experience. Of the last three companies I've worked with in the past three years, two of them have Diagnostics Pack licences and the other was Pythian. So, whilst Pythian's many customers might not pay for Diagnostics, the UK enterprises I've come across recently do - in fact that's a 100% adoption rate
2) I agree with Howard Rogers' assertion that Standard Edition customers should be allowed to licence the Diagnostics pack if they wish. In fact, Howard's written very perceptively about this over the course of the past year or so, if I recall correctly? However, if I had to make the choice personally, my stronger feeling is that this should be included in Enterprise Edition. If the product is truly an Enterprise product, then AWR is far more significant than any SOA/XML/Java/AQ blah, blah, blah feature. What do Oracle think that Enterprise DBAs *do*? Don't they think this stuff is central to our role?
3) In the end, I wouldn't mind so much if Oracle didn't give away so much software for free - SQL Developer and the rest. To restrict access to the workload repository just makes people suspicious - for what? It just makes the company look small-minded. When the licence restrictions become the target for humour - the law is an ass - it's time to reconsider.
I hope those minor differences don't detract from the central argument because I do worry that, the more fragmented the argument, the less likely it is to be won. I even accept the argument might be too emotional but never under-estimate the emotions of a bunch of geeks
Here's the thing, Larry. Wouldn't it be nice to be the good guy for once?
(Yeah, I know, it's business, blah, blah ... but make some fans happy and who knows how powerful that might be ...)
(Yep, time for another lie down ....)
#1 - Niall Litchfield said:
2007-07-04 21:17 - (Reply)
Of course, it's wrapped up in a site licence, but maybe that's what this is really about? A negotiating tool when dealing with big customers
That, in my view is exactly what it's about. RAC has the same issue (with the market distorting but it's free for SE and below customers thing). Site Licenses don't exist. Ask your rep ^H^H^H account manager. Now you know, I know, and indeed perhaps the world knows that site licenses *do* exist, but really only for 'large customers'. ISTM that the management packs are thrown in as freebies for people on a different deal anyway and left as threats for 'ordinary' customers.
Still I await with interest the 'I paid full price for everything' rebuttals.
Niall
who works for a rather large org that doesn't get the deals.
#1.1 - Doug Burns said:
2007-07-05 18:32 - (Reply)
Niall,
You're right and that's one of the unfortunate aspects of this saga. Of course Oracle are going to look after their big customers, but if they're 'chucking it in' as part of a site licence, they might get as much mileage from letting everyone use it.
>> Still I await with interest the 'I paid full price for everything' rebuttals.
You won't get any from me ![]()
#2 - Paul Vallee said:
2007-07-04 21:49 - (Reply)
I see you referring to 100% vs some smaller %, and I tried to determine the exact number of Pythian clients that are licensees of the Diagnostic Pack, and to the best of my abilities the number is two. Out of 70. And while it _may_ be a UK/USA thing, there really are some big customers in that list so I don't think it's a simple big shop small shop issue. While it's clearly inappropriate to discuss individual companies' licenses, since I can say that our active customer list includes AXA, Northrop Grumman, CVRD Inco, MDS, Fox Interactive, JDS Uniphase, Electronic Arts, Cinram, and Diageo, to a "big companies" only, so you can deduce that at most two of those are licensed for the environments we manage! I think this is pitiful market penetration for such a useful tool, and this also speaks to DBA's inability to convince senior (purchasing) management of their value. But that's not changing soon, is it? ![]()
Paul
#2.1 - Doug Burns said:
2007-07-05 18:41 - (Reply)
I see you referring to 100% vs some smaller %
To be clear, that was a response to the claim that
"To give a statistically significant example, of Pythian’s 70 outsourced DBA-for-Oracle customers, so few have licensed the use of this feature as to approach zero."
So I would never take my tiny sample as significant, but if I see contradictory evidence, I feel obliged to mention it.
Do you really think that 70 customers is a statistically significant example of Oracle's customers? I know a small sample can be remarkably accurate, but that seems extremely small to me and also skewed by the fact that they're one company's customers.
so you can deduce that at most two of those are licensed for the environments we manage!
I know you can't say much, but if the two customers are in that list, it would tend to reinforce the big site/small site thing.
I think this is pitiful market penetration for such a useful too
I agree, whatever version of the numbers we use, but it would be very interesting for Oracle to publish the real numbers!
and this also speaks to DBA's inability to convince senior (purchasing) management of their value.
I'll take that as a compliment to Enterprise DBAs, then
Seriously, perhaps it's easier for on-site DBAs to make the case?
#2.1.1 - Paul Vallee said:
2007-07-05 19:02 - (Reply)
Hi Doug,
Since we don't have a true random sample, sampling error as a result of a small sample is really the least of our worries. Factors likely to obliterate the confidence of the sample are more likely to include geographical selection, or even whether companies that choose to outsource some or all of their DBA work might be frugal including licensing rules. That being said, if there were no skew to the sample 70 observations is usually enough for your 95% confidence intervals to keep within 15% of your finding.
Cheers
Paul
#2.1.1.1 - Doug Burns said:
2007-07-05 19:05 - (Reply)
You've lost me, Paul!
No, I'm only kidding
I agree with your comments. Really, I was just rallyng against any attempt to use one group of customers to represent the whole.
#3 - Alex Gorbachev said:
2007-07-05 00:28 - (Reply)
First of all I would like to welcome the return of the good tradition! Randy, welcome to the "cover page"! Who did that shot? Just like a photo model. ![]()
If Oracle leaves Diagnostic Pack with "fancy" stuff - it would be still what it is now - "a negotiation tool" so Oracle would only win - users happy and nothing lost in licensing fees.
My subjective observation is that adoption is very low due to inaccessibility. Sometimes smaller customers go for it as well (I have one) but this is *very* rare.
#3.1 - Doug Burns said:
2007-07-05 08:49 - (Reply)
>> My subjective observation
Is that the same as a guess? ![]()
Alex/Paul,
Interesting comments both but I'm a bit late for work this morning
so will reply properly later.
#3.2 - Doug Burns said:
2007-07-05 18:47 - (Reply)
Alex, sorry, it wasn't fair to reply with the 'Guess' joke but I do promise you it was a joke. Anyway ...
Who did that shot? Just like a photo model. ;-)
It was me!!! Did you notice that one of his 'Hates' is 'guessing', although I must give Dan Fink credit for that ![]()
If Oracle leaves Diagnostic Pack with "fancy" stuff - it would be still what it is now - "a negotiation tool"
I agree.
My subjective observation is that adoption is very low due to inaccessibility.
Let me tell you something which is sadder ... Imagine if the companies who could afford this stuff had DBAs who didn't even know they were licenced to use it! That's no criticism of the DBAs, but when a company is big, it can take a while to find out that we can use it.
So, to reiterate, I'm sure it will sadden you to know that there are companies who have paid and don't use it at all
Whilst others would love to have it but can't.
To reiterate my comments in the blog above, for any disagreements on detail, I agree with any effort to improve adoption of ASH/AWR because I think they're fantastic tools, and I think Oracle would get a lot of good publicity from it.
#4 - joel garry said:
2007-07-06 01:24 - (Reply)
While googling for a cite for Larry's denigration of what DBA's do, I got distracted by this prediction of what dba's will be doing in 2007.
#4.1 - Doug Burns said:
2007-07-06 07:17 - (Reply)
Nice one, Joel, thanks for that ![]()
#5 - Noons said:
2007-07-06 11:03 - (Reply)
Of course I can't speak for other states in Australia. But here in NSW, I have yet to see a significant number of *prior* customers taking the awr plunge.
Mostly because they all have dbas who can easily achieve the same or have something similar already in place. Third party or homebrew. Hardly worth another investment, particularly in the current anti-IT, anti-Oracle climate over here.
I've seen quite a few takers in new clients. But nothing that I'd class as
significant - although of course, as usual, the local O marketing will disagree and claim 100% use.
Is it worth the moolah being asked for it? No way.
Larry claims dbas are expensive and their overhead should be reduced and then goes and makes a tool to help them in their job even more expensive? And the darn thing can't be switched off, but if we peek into it we gotta pay? The whole thing just doesn't make sense!
I know: it's not Larry. That's sooo not the point...
#5.1 - Doug Burns said:
2007-07-06 18:25 - (Reply)
>> Is it worth the moolah being asked for it? No way.
But you know big business, they'll often pay for it if they don't have something suitable in place already.
I'm also a bit of a fan of ASH/AWR to be honest but similar things can definitely be built.
I actually hope that you and Mark are right about the low adoption rates because Oracle must know the true rates and, if they're low enough, that's probably the best chance of them reconsidering.
I really enjoyed that SQL Server blog, btw. I'll link to it later.
#6 - John Hurley 2007-07-07 00:36 - (Reply)
I thought Niall had some kind of petition or something going?
Has there been any response to that officially?
Mark Townsend at one point said that he was going to respond on cdos but apparently backed off.
My site is apparently one of the few licensed for the diagnostics and tuning packs.
But there is a huge mess for Standard Edition customers.
It sure seems like oracle wants people to use it but not be able to pay for it.
It's a real mess.
#6.1 - Doug Burns said:
2007-07-07 08:24 - (Reply)
I thought Niall had some kind of petition or something going?
Has there been any response to that officially?
John, Niall's written more about the petition in his recent blog and Mark posted a follow-up blog that mentions it too. However it's fair to say that the petitions are different but complimentary. Both have had responses (of a kind) from Oracle employees. Whatever gets Oracle thinking about the issue is useful in my view.
But there is a huge mess for Standard Edition customers.
I agree because they can't licence it at all, as Niall HJR and others have pointed out. I'm not typically working with Standard Edition so I've not been burnt by that one as much.
It's a real mess.
It's certainly that. The fact that I've head Oracle employees are often as confused as we are by the licencing is an indication of that.
Another example is the ludicrous situation of the lack of a supported method to switch it off, without infringing the licence. It looks like they're addressing that one, but it ws always laughable.


Sorry, I shouldn't be so sarcastic, but try to show some sympathy for my schedule.Thursday 9th August 22:30 BST - Got to bed, unusually early.Friday 10th 06:00 BST - Wake up, check Netvibes and noticed several 11g release blogs, including Eddie's initia
Tracked: Aug 11, 15:06