I had a little exchange with someone on Twitter last night, which was initiated by him complaining about the cost of Oracle and predicting their demise. Once that was over I spent a little time thinking about my “fanboy status”.
If you know anything about me, you will know I’m an Oracle fanboy. I’ve spent nearly 20 years doing this stuff and the last 14+ years writing about it on the internet. If I wasn’t into it, it would be a pretty sorry state of affairs. So does that mean I’m totally blinded like all those Apple fanboys and fangirls? No. I just don’t choose to dwell on a lot of the negative and instead focus on the positive, like the cool bits of tech. The common topics I hear are:
- Oracle costs too much : I could bleat on about the cost of Oracle and what features are missing from specific editions, but quite frankly that is boring. Unless you’ve been under a rock for the last 35+ years you should know the score. If it’s got the name Oracle associated with it, it’s probably going to be really expensive. That’s why people’s jaws drop when they find out Oracle Linux is free. They are just not used to hearing the words Oracle and free in the same sentence. If you want free or cheap, you can find it. What people often don’t consider is total cost of ownership. Nothing is ever free. The money just gets directed in different ways.
- The cheap/free RDBMS products will kill Oracle : This talk has been going on since I started working with Oracle 20 years ago. It used to worry me. It doesn’t any more. So far it hasn’t materialized. Sure, different products have eaten into the market share somewhat and I’m sure that will continue to happen, but having a headstart over the competition can sometimes be a significant advantage. I work with other RDBMS products as well and it is sometimes infuriating how much is missing. I’m not talking about those headline Oracle features that 3 people in the world use. I’m talking about really simple stuff that is missing that makes being a DBA a total pain in the ass. Typically, these gaps have to be filled in by separate products or tools, which just complicates your environment.
- It’s just a bit bucket : If your company is just using the database as a bit bucket and you do all the “cool” stuff in the middle tier, then Oracle databases are probably not the way to go for you. Your intellectual and financial focus will be on the middle tier. Good luck!
- But company X use product Y, not Oracle : I’m so bored of this type of argument. Facebook use MySQL and PHP. Yes, but they wrote their own source code transformer (HipHop) to turn PHP into C++ and they use so much stuff in front of MySQL (like Memcached) that they could probably do what they do on top of flat files. Companies talk about their cool stuff and what makes them different. They are not so quick to talk about what is sitting behind the ERP that is running their business…
- NoSQL/Hadoop/Document Stores will kill RDBMS : Have you ever had a real job in industry? Have you ever done anything other than try to write a twitter rip-off in Ruby for your school project? Do you know how long it took COBOL to die? (it still isn’t dead by the way). There is a massive investment in the I.T. industry around relational databases. I’m not saying they are the perfect solution, but they aren’t going anywhere in the near future. Good luck running your ERP on any of these non-RDBMS data stores! What has changed is that people now realise RDBMS is not the right solution for every type of data store. Using the right product for the right job is a good thing. There are still plenty of jobs where an RDBMS is the right tool.
- The cloud will kill Oracle : The cloud could prove to be the biggest spanner in the works for many IT companies. If we start using cloud-based services for everything in the Software as a Service (SaaS) model, who cares what technology sits behind it? Provided our applications work and they meet our SLAs, who cares how many bodies are running around like headless chickens in the background to keep the thing running? For Platform as a Service (PaaS) and Infrastructure as a Service (IaaS), I don’t think cloud makes so much of a difference. In these cases, you are still picking the type of database or the type of OS you need. They are not hidden from you like in the SaaS model. I guess the impact of cloud will depend on your definition of cloud and route the market eventually takes. What people also seem to forget is the big winners in the cloud game will be the big companies. When the world is only using SaaS, you are going to have to work for Amazon, Oracle, Microsoft etc. if you want to be a techie. The ultimate goal of cloud is consolidation and centralisation, so you will have to work for one of these big players if you want to be anything other than a user. I find it interesting that people are betting on the cloud as a way of punishing the big companies, when actually it is likely to help them and put us folks out of business…
The post has got a bit long an tedious, so I’m going to sign off now.
In conclusion, yes I’m a fanboy, but I’m not oblivious to what’s going on outside Oracle. I like playing with the tech and I try to look on the positive side where my job-related technology is concerned. If I focussed on the negative I would have to assume that Oracle is doomed and we will all die of Ebola by the end of the week…
Cheers
Tim…
“Oracle costs too much”.
I’ve been with Oracle since 7.1. and I say that it’s not only about price. It’s about that what are you getting for that price. I can state that quality of products became worse than it used to be. Quality of support became much worse than it used to be.
Now when I consider TCO (which I allways do because I have to, because there should be positive business case for such purchase) I say it is not problem with CAPEX, it’s a problem what you are getting for OPEX which you spend every year.
I’ve opened only 3 SRs within last 4 years. It is not because Oracle technology has best and no bug code written but it’s because I’d rather spend some tome with finding the workaround which usually takes much smaller portion of time then SR resolution from Oracle side.
First SR had severity 1 (escalated), I had been waiting for patch for 3/4 year until they released the patch according to all RDAs, traces and testcases provided during SR creation. Rest of the time I was responding to the same repetitive questions from “analyst”. Fortunately I found the workaround which prevented the instance crash every 3 days. This is not acceptable for any database nor DWH.
Second SR, my database was starting and performing very poorly. Again collected all possible traces and information (RDA, AWR or Statspack, oradebug, truss, and specific traces) and provided during SR creation. The same scenario Severity 1 (later escalated), stupid question and meaningless action plans (probably to satisfy internal Oracle KPIs). Performance was very really very bad and pretty not acceptable for the billing system for one month. So I checked the traces (uploaded during SR creation), read the HP-UX docs and created a workaround (not so simple) to overcome the problem. It was clearly bug in Oracle code which was (finally) fixed two PSUs later.
I’m not going to write about third SR, because it would be the same story. There is only one difference, affected system was CRM. Three SR for three very important business systems, with the same result – Oracle support failed. I’ve always logged SR only when it was really neccessary, I’ve tried to be sure that problem is
not on my side, but after this experience I’m trying to find workarounds instead of logging an SR.
We are large company and Oracle knows it. Our OPEX regarding Oracle technology spending is really interesting amount of money (because we have a lot of their stuff). And we know that and Oracle knows that as well. This is the reason why they are contacting and visiting us more frequently than Mormon guys in order to check the oportunities and to sell more (mostly useless) stuff to us. What we need is strong support services from quality point of view. No, we are not going to buy anything, instead of that we are decomissioning and reducing Oracle stuff and we want keep it at minimum required level in order to make OPEX (for Oracle) significantly lower. That’s why we don’t buy SPARC servers anymore, no Exadata anymore, no more licenses. That’s why we are adopting new thing from other vendors because they have the same speed or they even faster, modern with much better support and also fair price. And this escape from Oracle is worth from TCO point of view.
When I buy a Mercedes, I expect reasonable quality and reasonable support for that price. Quality of Oracle products and Oracle support lacks in this cases.
We know it, other customers know it, Oracle knows it and the Stock market knows it as well. Larry as freshly donwgraded CTO (as a result of Stock market news) hopefully can improve quality of products and support as well, to be trusted partner-vendor worth every cent again…
Till that time my questions will be still: Why I’m spending so much money (OPEX) for several SR(s) or do I have to spend so much money in the future?
Because these questions come to mind mind when I’m calculating TCO during comparation of several solutions. And yes, Oracle cost too much in most of the cases…
One addition. I know you, your past work and you are very brainy person but since last couple of years you became Oracle doctrine submissive person just like most of the rest of ACE Directors. When you were just ACEs, you were more trustworthy than now and you weren’t just Oracle interpreters… And unfortunately I’m not the only person with this opinion, but probably I’m only one who is not affraid to make it public (even hidden behind “noname” nickname).
noname : I understand many of your points and I think we have all been in this position at some points in our career. I’ve certainly worked in several companies where Oracle was not a good fit, mostly because of cost, yet it was still picked because the marketing hype fooled people into thinking it was the right tool for the job. We have all been at the mercy of Oracle Support being unresponsive also, so I’m with you there.
My only concern about your post was actually the last paragraph. I am disappointed that you would think my opinion has in any way changed as a result of being an Oracle ACE [Director]. I think this warrants a specific blog post with my thoughts on it. I’ll follow this up when I get home… 🙂
Cheers
Tim…