Oracle plugs socket numbers on DIY Standard Edition

If you are not already tied in

$17,500 ~= £11,500

So let's have four of those = £46,000

That buys you a three node VMware cluster on say Dell PE7x0 and an Equallogic iSCSI SAN plus a couple of reasonable switches. If you ditch the VMware and fire up containers on Linux (some assembly needed) you get a shit load of Postgres on that lot or simply run a three node cluster of DBs. There's also MariaDB and a whole heap of NoSQL to play with as well. All well supported and lots of integrations available.

The savings on Oracle? Buy some consultancy ...

Remember the first rule of the Oracle club

Is that under no circunstances with each new release you will pay less for the privilege of running exactly the same piece of software you are running today.

By the way, how have they gone so far as to be able to charge customers for upgrades (that rarely are ever applied) and support (that unless you are a massive customer is close to not having any or even worse) AND at the same time reserve the rights to change licencing T&C's as part of the upgrade? Or is there something that I've misunderstood?

go Postgres, go

Seriously, for lower-end stuff definitely a good alternative. Serious SQL, not just some lashed-together webby db.

And tell you what, I dunno exactly how well their Oracle compatibility mode works, but my guess is that, if you play your cards right and watch both your code and application architecture, you could potentially upgrade to real Oracle when your customer base justifies it. And I still think Oracle is a wise database choice, if you have real reasons to afford it.

IMHO, Oracle should be slightly more careful. I don't know what their selling point is in the context of say AWS or Digital Ocean. I know you can run SQL Server on AWS. I also know that Chef-ing just an Oracle client on a Vagrant/Virtualbox is already somewhat more cumbersome than strictly necessary.

Cutting yourself off from trendy DIY starving artist kinda cloudy deployments may or may not be a wise strategy, time will tell. Customers who don't pay anything are not very useful, but neither are belatedly lucrative ones that went to a competitor.

Where is the database competition on *Nix boxes

It is too bad there is so little competition in commercial RDBMS alternatives. That is why Oracle can ask what it wants, nobody will use MySQL for an environment where an outage costs 6 digits or more per hour, although MySQL/MariaDB is very good, never see it crash even in heavily used E-commerce websites.

Windows = SQL Server or Oracle, where most Windows shops of course go with SQL server

Unix = Oracle and DB2 from IBM

Oracle RDBMS is just sad. The tooling they offer for database management is an insult to the user, even when being a *Nix fanboi, i rather use Management Studio MS provides for Sql Server than this oracle stuff.

DB2 from IBM is unfortunately so much underestimated. IBM provides this with its WebSphere E-commerce suite, and it works like a dream. Offering a good and comprehensible command line interface, something oracle totally lacks. Too bad people ignore it for decades now, if they would buy it more, Oracle prices would be lower and its products better.

You may not get fired

You may not get fired for buying Oracle, but these days you won't win the Mr.Popularity competition either. UK government loathes Oracle, and quoting them in a proposal is the kiss of death in some departments. In fact so much so, that they will do anything to get away from Oracle's ever increasing support costs, including trashing major systems, or spending millions on converting to another vendor.

Postgres is becoming a lot more popular as there is a low cost conversion path for PL/SQL.

Somehow Oracle have lost the plot with products that are now too expensive, they may have a really good Enterprise DB, but it has capabilities that really aren't needed for most applications, where as the Standard Editions really are good enough for most applications, so it's no surprise that people see it as a great alternative. Hobbling it even more can only loose them business in the longer term, why would you buy SE when you will face an exponential jump in costs as your business grows? There are other alternative out there and SMEs know who they are.

It's a long time Oracle tries to fight CPU evolution...

... and usually it loses.

A few years ago it tried to charge you by the MHz (back then CPUs were still clocked in a few hundred MHz). In a few months, CPU "accelerated" so fast Oracle prices became astronomical - and when you were forced to buy a new CPU because the old system was dead, you had to pay more just because the older, slower, CPU was no longer available.

Then when multicore CPU arrived, it tried a complex mechanism given each core a weight depending on the architecture - understanding how much an Oracle instance would cost you required careful reading of all the rules.

Now again it's trying to limit what the software you paid for achieve when you upgrade the hardware.

Even if I agree that for "standard" edition a limitation on the number of "socket" supported, and cluster nodes) may be acceptable - systems with more than two sockets and large cluster are pretty expensive hardware too - I really don't understand a cap on the number of threads, etc. for the single CPU - it's just the normal evolution of hardware, something Oracle can't accept as more "entry level" (for Oracle) systems becomes capable of handling larger workloads at a cheaper price (that's plain IT evolution, Larry...) and thereby less customers needs the more expensive products.

A more clever company would try to broaden the customer base, instead of trying to blackmail actual customers... not Oracle. But not always works, it may be forced to back off.

socket pricing

As in "socket to the suckers"?

Mind you, I like the Oracle database, but the licensing is a bear.