The announcement that several Oracle products will be released “Cloud First” from now on has been an interesting talking point for many of us. I think I first heard this message at last years OOW, but I can’t remember if it was public, or in a meeting covered by an NDA. There have been public statements about this since then, but at OOW16 we have the first example that directly affect me!
Oracle Database 12c Release 2 is here, kind-of. If you sign up for the “Oracle Database Exadata Express Cloud Service” you will get a PDB on Oracle 12.2. So what are my thoughts on Cloud First?
Conspiracy: “Oracle are forcing us on to their cloud. Won’t someone think of the children?” Come on. Let’s be real. I speak to people all the time who have yet to get all their systems to Oracle 11g. Making them wait a bit longer for 12.2 is not going to be a big deal for most businesses.
Bugs: The Oracle database is a big and complicated product. It is inevitable that new versions will contain bugs. The most important point is how long will it take to fix them? In the past we have had the, “wait for the second release”, and more recently the, “wait for the first patch set”, mindset. I can understand the later. The 12.1.0.1 release was very buggy. The 12.1.0.2 release felt solid. If it is done *properly*, Cloud First will allow Oracle to patch bugs really quickly, so most users never have to encounter them **. By the time we get an on-premise release, it should be solid. Hands up how many people were going to go live with 12.2.0.1 soon after release date? None of you? Yeah. I guessed that. Perhaps when we get the first on-premise release of 12.2 it will actually be production ready!
Quicker release cycle? : For many of the products, the Cloud First approach means they will be able to do a quarterly release cycle, with a yearly on-premise release. I think that’s a good thing. Will it result in the same release cycle for the database? I highly doubt that, but it would be nice if we get releases a bit quicker. A new release every 3-4 years is getting out of hand…
Evangelism: This is one that hits me hard. I’ve not been on the 12.2 beta program. I’ve not had access to the product. Now 12.2 is available I can only get hold of it via a cloud subscription and even then it is fully managed, so I can’t practice many of the DBA tasks on it.
I’m guessing most of the beta testers who have prewritten articles waiting to go live can’t actually release them until the on-premise release appears. If they do, they are not based on the production release. If they are DBA articles, they definintely can’t be released as there is no way they can test DBA stuff on the current production release and releasing articles based on a beta product is doing you a disservice and may still be breaking their NDA.
Without wanting to sound overly dramatic, this means the 12.2 release is “dead to me” until there is an on-premise release. So Oracle should panic because Tim Hall isn’t happy about something right? Of course not. I think evangelists like myself will be affected, but so what? Those articles will still get written ***. They will just arrive a few weeks/months/years later. It’s annoying for me, but it’s not going to bring Oracle down. 🙂
Conclusion : Although the Cloud First approach to product delivery is extremely annoying to me personally, and will keep the conspiracy theorists busy, I can see some really positive sides to it if it’s done correctly. Time will tell!
Cheers
Tim…
Updates: Based on comments.
- ** Of course, someone has to identify the bugs, which is why I said “most users never have to encounter them”. Many companies run with un-patched systems for a long time and encounter bugs that were fixed a long time ago. In an ideal scenario, one customer finds the bug, it gets fixed and everyone is patched preemptively. 🙂
- *** What about consultants? Until the on-premise release is available, nobody has access to the on-premise product. As a result, you are in the same position as everyone else that was not on the beta. Early adopters, like consultants, will start their early adoption from the point where the on-premise release is available. Who is going to ask you to do DBA work on the version before that? The only access a customer can have is the fully managed system, so nobody. If you *must* be working with 12.2 from day one for fear of being left behind, pay your $170 per month… Judging by the typical adoption timescales, this is not an issue for the vast majority of consultants.
Hey Tim,
> If it is done *properly*, Cloud First will allow Oracle to patch bugs really quickly, so most users never have to encounter them. By the time we get an on-premise release, it should be solid.
Really? Who should find all of these bugs? I mean Oracle should patch bugs really quickly just because they are running it in the cloud on their own? The cloud is a standardized deployment and so they do not run edge cases like customers do in their environments. Otherwise the “standard” bugs should already be found in beta testing and development phase.
> So Oracle should panic because Tim Hall isn’t happy about something right? Of course not.
No, but they f*** up all the independent consultants and third party companies who implement and support their product. I mean it is extremely important to be “early adopters” for us (independent consultants) as we need experience and in-depth knowledge with the product when the clients are ready to use it. Our knowledge advance is the critical business part here … delay is the problem.
Just my 2 cents.
Regards
Stefan
P.S.: I also got no chance to get into the beta program (just because of no own CSI) and cloud access is also not possible (http://blog.dbi-services.com/question-is-upgrade-now-to-12-1-0-2-or-wait-for-12-2/#comment-3859). That is ridiculous.
Stefan:
Bugs: Many people run non-patched systems for a long time, so they encounter bugs that were fixed a long time ago. Of course someone has to identify a bug, but in the best case that might be one person, the bug gets fixed, everyone is patched and nobody else suffers. Much better than everyone having to live through the same pain…
Consultants: Yes, that is true, but you still have options. They only thing available today is the fully managed service, which you can buy for $170 per month if you feel you must have access to it. I don’t care enough to pay. 🙂 Until the on-premise release is available, you will not have any need for the other skills. Once the on-premise release is available, you will still be an early adopter like the rest of us first-timers. I don’t see the difference.
Beta: I think a lot of the time access to the beta program is overrated. Very few people push the products hard. Most of it is really basic functional testing. You never know what is going to make it to the final production release. I’ve kind-of given up caring about betas. Better to wait for the production release, which in this case is the on-premise release, not the Cloud release…
Cheers
Tim…
Hey Tim,
> They only thing available today is the fully managed service, which you can buy for $170 per month if you feel you must have access to it. I don’t care enough to pay.
I really don’t care much about the price, but PDB access only is useless for me. I don’t want to deploy a schema and write some fancy SQL for app data. I want to dump/trace stuff and i wanna have a look under the hood how new features really work 🙂
> Until the on-premise release is available, you will not have any need for the other skills. Once the on-premise release is available, you will still be an early adopter like the rest of us first-timers. I don’t see the difference.
Well the main difference is the time gap between releasing the “full” 12.2 on-premise and when clients are considering to use it. Let’s say Oracle would publish the “full” on-premise 12.2 release right now and clients want to use it in 4 month or so (due to 11.2 support restrictions and the need to upgrade) – then i would have round about 4 months for research. But as Oracle puts the GA release of 12.2 closer to the 11.2 support restriction – less time or maybe none at all for research and providing clients the corresponding consulting value for the money.
Regards
Stefan
Stefan:
I feel your pain, but as I said before, we are all in the same boat. I woudn’t trust anyone claiming to be a 12.2 expert for a long time. 🙂
Cheers
Tim…
Interesting marketing concept, this one:
“Don’t let anyone use the product until all the bugs are ironed out”?
Seriously?
I’m still installing PSU patches for 12.1.0.2 and the patching itself still has bugs!
Amazing how anyone at Oracle could come up with such a “brilliant” concept… 😁
This is the approach Microsoft has been taking for the past 6-8 years, since they started building Azure. I think it has worked out well for them. I’m kinda shocked that Oracle has taken so long to follow suit. As an outsider, I’ve never known Oracle to play catch-up to anyone else, they are usually playing with the lead and not from behind. Will be interesting to see if they can get caught up and help push everyone to (whatever)-as-a-service model.
Thomas: I think it has taken them this long because they came to the cloud game late. Much later than Microsoft. 🙂
Cheers
Tim…