Unconventional Wisdom: Do we probably need RAC, Exadata, Oracle Database 12c, MySQL, certification, and Oracle user groups?
Years ago, Mogens Norgaard, the co-founder of the Oak Table network wrote a provocative paper titled “You Probably Don’t Need RAC.” Here are the opening sentences of his paper:
If you’ve been holidaying in Siberia or similar places for about a year, you have probably not talked to an Oracle Sales rep yet about RAC. But you will no doubt find that there’s a voice mail waiting for you when you turn your mobile phone on again after returning home from the vacation.
RAC is being pushed very hard by Oracle. You will get high availability, incredible scalability, a much improved personal life, the ability to partition workloads, buy cheap Linux servers and what have you.
It sounds pretty good. How can anyone say no to that kind of offer?
The closing sentences are as interesting as the opening sentences, especially the very last one.
If you have a system that needs to be up and running a few seconds after a crash, you probably need RAC.
If you cannot buy a big enough system to deliver the CPU power and or memory you crave, you probably need RAC.
If you need to cover your behind politically in your organisation, you can choose to buy clusters, Oracle, RAC and what have you, and then you can safely say: “We’ve bought the most expensive equipment known to man. It cannot possibly be our fault if something goes wrong or the system goes down”.
Otherwise, you probably don’t need RAC. Alternatives will usually be cheaper, easier to manage and quite sufficient.
Now please prove me wrong.
The paper was written in the early days of RAC. The technology has matured and improved since the date of the paper and, therefore, a number of the technical details in the paper are no longer valid. However, the underlying message of the paper is that you need to make an informed decision, justify the increased complexity and cost, and consider the alternatives.
Years ago you said that we probably don’t need RAC. Have you recanted yet? Do we probably need RAC?
I still think very, very few shops actually need RAC. Fantastic technology—just like, say, head-up display (HUD) for cars—but few really need it. RAC still has all the hallmarks of something people will want to buy: It increases complexity immensely, it’s expensive, it requires specialists that are increasingly hard to find, there are always excellent alternatives—and it’s pretty much perpetually unstable. For all those good reasons, more and more customers are using it. Either because manly types like to increase chaos, or because I’ve been telling people not to use it since around the year 2000. Whenever I recommend or don’t recommend something, most customers go out and do exactly the opposite, so in that sense I have a great deal of influence in the market.
Whenever I recommend or don’t recommend something, most customers go out and do exactly the opposite, so in that sense I have a great deal of influence in the market.
Do we probably need Exadata? Is Big Iron the ultimate answer to the great question of life, the universe, and everything?
In some ways, Exadata is the new RAC. It’s a lot about hardware, uptime, performance, amazing technology—and price. It’s also approaching the “Peak of Inflated Expectations” as seen in Gartner’s hype cycle, and it will soon set its course downwards toward the “Trough of Disillusionment.” Just like with RAC, I simply love the technology—a lot of good guys that I like and respect are on it, but few really need it. One of the things I love about it is that there isn’t any SAN involved, since I believe SANs are responsible for a lot of the instability we see in IT systems today. I tend to think about Exadata as a big mainframe that could potentially do away with hundreds of smaller servers and disk systems, which appeals hugely to me. On the other hand, the pricing and complexity makes it something akin to RAC—that’s my current thinking.
One of the things I love about [Exadata] is that there isn’t any SAN involved, since I believe SANs are responsible for a lot of the instability we see in IT systems today.
Do we probably need Oracle Database 12c (or whatever the next version of Oracle Database will be named)?
Since Oracle 7.3, that fantastic database has had pretty much everything normal customers need. It has become more and more fantastic; it has amazing features that are light years ahead of competitors—and fewer and fewer are using the database as it should be used (they’re using it as a data dump, as Tom Kyte said many years ago), so the irony is that as the database approaches a state of nirvana (stability, scalability, predictability, diagnosability, and so forth—fewer and fewer are using it as it should be used (in my view), and more and more are just dumping data into it and fetching it.[Norgaard’s First Law]
Since Oracle 7.3, that fantastic database has had pretty much everything normal customers need.
Do we probably need MySQL? Or do we get what we pay for?
As customers (and especially new, freshly faced programmers) want to use new things instead of things that work and perform, it becomes more and more logical to use MySQL or other databases instead of the best one of them all: Oracle. Since MySQL succeeded in becoming popular among students and their professors, it is immensely popular among them when they leave school (the professors stay, of course, since they don’t know enough about databases to actually get a real job working with them outside academia). So MySQL will be used a lot. And it’s an OK database, especially if we’re talking the InnoDB engine.
As customers (and especially new, freshly faced programmers) want to use new things instead of things that work and perform, it becomes more and more logical to use MySQL or other databases instead of the best one of them all: Oracle.
Do we probably need certification? Or do we learn best by making terrible mistakes on expensive production systems?
I hate certifications. They prove nothing, and they become a very bad replacement for real education, training, and knowledge. Among Windows and Cisco folks, it’s immensely popular, but you can now feed all the farm animals in Denmark (and we’ve got quite a few, especially a lot of pigs) with certified Microsoft and Cisco people. It’s taken by students (what?!? instead of real education, they train them in something that concrete? I find it really stupid), among unemployed (we have a lot of programs for those folks here), and what have you. They’re worthless, and a lot of people think it will help them finding a job, thereby providing false hopes and security. YPDNC.
I hate certifications. They prove nothing, and they become a very bad replacement for real education, training, and knowledge.
Do we probably need ITIL? Should we resist those who try to control and hinder us?
When you begin doing “best practices” stuff like ITIL, you’ve lost. You’re pouring cement down the org chart in your shop, and God bless you for that—it helps the rest of us compete. “Best practices” means copying and imitating others that have shops that are unlike yours. Standardizing and automating activity in brain-based shops always seemed strange to me. The results—surprise!—are totally predictable: jobs become immensely boring, response times become horrible, queues are everywhere, and nothing new can happen unless a boss high up really demands it. It’s Eastern Europe—now with computers. Oh, and it’s hype; it’s modern right now but will be replaced by the next silly thing (like LEAN—what a fantastically stupid idea, too). Maybe we’ll have LEAN ITIL one day? Or Balanced Score Card–adjusted ITIL? Or Total Quality Management of LEAN ITIL?
When you begin doing “best practices” stuff like ITIL, you’ve lost. You’re pouring cement down the org chart in your shop, and God bless you for that—it helps the rest of us compete.
The funny thing is that Taylor’s ideas (called “scientific management”) were never proved, and he was actually fired from Bethlehem Steel after his idiotic idea of having a Very Big Hungarian lift 16 tons in one day (hence all the songs about 16 tons), because he cheated with the results and didn’t get anything done that worked. Not one piece of his work has ever been proved to actually work. His “opponent” was Mayo (around the 1920s), with his experiments into altering the work environment (hence the constant org changes and office redos that everybody thinks must be good for something)—and his work has never been proved either. And he cheated too, by the way, which he later had to admit. So all this management stuff is bollocks, and ITIL is one of its latest fads. I say: Out with it. Let’s have our lives and dignities back, please.
NoCOUG membership and attendance has been declining for years. Do we probably need NoCOUG anymore? We’ll celebrate our 25th anniversary in November. Should we have a big party and close up the shop? Or should we keep marching for another 25 years?
No. Oracle User Groups are dead as such. Just like user groups for mainframe operators or typesetters. You can make the downward sloping angle less steep by doing all sorts of things, but it’s the same with all Oracle user groups around the world. I think I have a “technical fix” or at least something crazy and funny that can prolong NoCOUG’s life artificially: move onto the Net aggressively and do it with video everywhere. Let it be possible to leave video responses to technical questions (why doesn’t Facebook have that?); let it be possible to upload video or audio or text replies to debates and other things via a smartphone app. Let there be places where the members can drink different beers at the same time and chat about it (and show the beer on the screen), etc., etc. In other words: Abandon the real world before all the other user groups do it—and perhaps that way you can swallow the other user groups around you and gradually have World Dominance.
Oracle User Groups are dead as such. Just like user groups for mainframe operators or typesetters. You can make the downward sloping angle less steep by doing all sorts of things, but it’s the same with all Oracle user groups around the world.
It costs a fortune to produce and print the NoCOUG Journal. Do we probably need the NoCOUG Journal anymore?
I have subscribed to the world’s arguably best magazine, The Economist, since 1983. Recently they came out with an app, and now I don’t open the printed edition any more (I still receive it for some reason). It’s so much cooler to have the magazine with me everywhere I go, and I can sit in the bathroom and get half of the articles in there read. It’s the way. Magazines should not be available anymore in print. Nor should they (in my view) be available on a silly website that people have to go to using a PC, a browser, and all sorts of other old-days technology. The smartphone is the computer now. Move the magazine there aggressively, and in the process, why not create a template that other user groups could take advantage of? Or the Mother of All Usergroup Apps (MOAUA) that will allow one user group after another to plug in, so people can read all the good stuff all over the world?
Magazines should not be available anymore in print. Nor should they (in my view) be available on a silly website that people have to go to using a PC, a browser, and all sorts of other old-days technology.
I’m writing a book on physical database design techniques like indexing, clustering, partitioning, and materialization. Do we probably need YABB (Yet Another Big Book)?
No, certainly not. Drop the project immediately, unless you can use it as an excuse to get away from the family now and then. Or, if you must get all this knowledge you have out of your system, make an app that people can have on their phone and actually USE in real-life situations. Abandon books immediately, especially the physical ones.
Interview conducted by Iggy Fernandez for the August 2011 issue of the NoCOUG Journal.
BACK TO POST Footnote: Norgaard’s First Law is “Whenever something reaches a state of perfection, i.e., where it becomes stable and very productive (in other words, saves time and money and effort), it will be replaced by something more chaotic.” (Q2 2006 issue of the ODTUG Technical Journal).