Microsoft hikes SQL Server 2012 price 'by 20 per cent' Microsoft will hike licensing prices for the soon-to-launch SQL Server 2012 Enterprise edition by an average of 20 per cent, resellers have claimed. From, er, April Fools' Day, customers who buy the top level version will be charged based on the numbers of cores licensed rather than the number of servers used. The Standard …
From the purely market-side perspective, this price hike is barely justifiable. There are not much really ground-breaking new features to talk about in the 2012 version. You must be insane to be willing to pay that much money to deploy Hadoop, which is of course quite possible :- )
From the perspective of reve-generation, it is fair to say Microsoft has probably finally seen the light. To wit, if you add features to scale a SQL Server Datacentrer to Terabytes of RAM and use in-RAM processing which is one of 2008 R2 features (+ database partitioning coupled with multicore), you do get really nice speeds for any workload. Why buy more gear and licenses, then?
Add to that that with bigger datasets traditional disks are not in fashion anymore and the I/O you get from Flash, in ant form, prevent you from adding more server power and the SQL Server licenses which inevitaly follow.
Microsoft is therefore the victim of the success of its own technologies and is now trying to regroup, to prevent that from showing in the bottom line.
One can still easily notice, too, that any price hike Microsoft devises in the forseeable future won't even touch the prices paid for Oracle, even after "larry's discounts".
A client was using one of my applications which uses a Firebird back end. They asked me to write a bespoke application for them but company policy was to use SQL Server. 2 years later, they asked me to switch the new app to Firebird. It is robust, reliable and fast. The fact that it is free was an issue but only in the sense that they could get for nothing what Microsoft charge and arm and leg for.
Why pay Microsoft these inflated licence fees?
btw As a member of the the Firebird foundation I pay to use it.
Plenty of reasons, though you wouldn't think of them if you're a small company and the only thing you care about is price. If I was an IT manager at a large company, this is what I'd be concerned about:
1. SQL Server has a proven track record. I've never even heard of Firebird until now, so I would have to do a lot of research and a lot of testing before putting it to real use.
2. Are there benchmarks comparing performance? Firebird is cross-platform and that *usually* means at least some hit to performance. Plus, I'm sure they perform differently with different tasks so performance-wise one might not be a right fit for what I'm doing.
3. Have any of my employees ever used Firebird? How difficult will it be for them to learn a new system? Keep in mind if I have to spend $30,000 training my employees on Firebird, I'm not really saving money.
4. How many new hires will have experience with Firebird? Tons of websites are built with PHP because tons of people have PHP experience. Far fewer are built with Ruby or ColdFusion because if your programmer leaves your hiring pool is limited.
5. How secure is Firebird? Can I hold the Firebird team accountable if we have a breach of security? Security bugs in SQL Server would be a huge problem for MS because they need to make money and no one will buy the product if it's not secure. A couple of open source developers don't have a business incentive to care.
6. Will Firebird be around in three years? Let's say the product is really good, but instead of eating Ramen for the rest of their lives the open source developers want to earn some money, abandon the project, and leave me high and dry.
And I'm not even an IT manager. I imagine someone that does this for a living could give you a dozen more reasons. And FYI, all open source software claims to be "robust, reliable and fast" (i.e. Android, which I've found to be rickety, unreliable and slow).
MS SQL has a few more features than Firebird (SSIS, XML, a few additional data types), but they are in the same league. Firebird is a MySQL alternative which is an MS SQL alternative. If you are concerned about performance, security, enterprise grade functionality, you are using Oracle or DB2, not MS SQL or Firebird.
You really should check out Firebird. As far as track record is concerned, it started life back around 1982/83, long before SQL Server was though of. It has existed in different incarnations, notably InterBase which is still marketed by Embarcadero.
There are plenty of large, mission critical installations of Firebird around the globe. But no cost means no marketing. Do not be blinded by the Microsoft marketing machine.
As for support, I've never needed in the 14 years I've used it. But there are companies who do support and they have close links to the development team.
1. The only proven track record I know about concerning SQL Server is that it beats MS Access performance-wise ... apart from that, it is a pile of crap. Even Sybase does a better job and they use the same code-base.
2. Firebird is cross-platform and that *usually* means at least some hit to performance. - No shit, man, you "are" funny ... Nothing actually beats UNIX performance-wise, because you shut off/remove anything you do not need running with ease ... unlike Windows where you even have the OS/2-emulation layer running just in case you are one of the two companies in the world running mission-critical OS/2 programs on Windows .... ROFL, ROFL^10 - fuck, I am really gonna wet myself
3. Databases is pretty straight forward, when you know one well it can hardly be very difficult to learn another over a weekend ...
4. See point 3. - Or are you hiring useless paper MCSE's who don't know what cmd.exe is ?
5. If you need a really secure RDBMS, then MS SQL simply cannot be your choice, seriously ... look at Oracle or DB2 ...
The issue is generally one of support. If the shit hits the fan and you can get vendor experts on the ground helping you etc it passes enterprise operational risk requirements. If the support base is smaller/less well known/less available etc it gets flagged as a risk. Like it or not that's how it works. Get too risky a rating on the OR checklist and you won't be allowed to proceed. Also boards/trustees do not like reading external audit reports highlighting their use of unconventional technologies as it makes the old farts nervous.
I have been hearing that while the change is to a "per core" license, the actual cost for 4 cores would be the same as the original "per CPU" license. Which is find for people with quad core processors, however with the change to hex core and more, yeah I can see where the prices will go up.
Still, its just the thing these days, license costs just keep going up.
Yep, per core licensing (or Server + CAL - the CALs are going up 20ish% from 2k8 prices).
Makes sense though, with the increasing core per cpu count, MS's per processor licensing was getting increasingly generous to the customer. Reached the point now where they'd be mugging themselves on prices!
when evaluating MS SQL vs Oracle for small-scale applicaions - particularly BI systems. The simple licencing model from MS and easily recognised value for money, especially when compared to alternatives, made choosing MS SQL a straightforward decision that required little research and basic business case. The proposed (new) pricing will make me (and others) look more carefully at price and what's included and look closely again at Business Objects/Cognos etc or Open Source options.
Some customers will see the price go up much more than 20%.
Many enterprise customers' standard box is a 4x10 server--four procs, 10 cores each.
With SQL Server 2008, they could license the entire box using SQL Server Enterprise, with four proc licenses costing about $25,000 each, or $100,000 for the whole box. To do the same thing with SQL 2012, they need to license 40 cores, at about $7K each, which will cost them $280,000. Then, they need to have Software Assurance as well (which wasn't required with SQL 2008 if you wanted unlimited SQL VMs). That adds 75% to the price of the product over three years, and continues after that at 25% a year. So that's another $210,000 for the first three years. Total: $470K--nearly a half-million to license one of these boxes if you need to buy SQL licenses for it and want the right to run unlimited SQL instances. A bit more than a 20% increase, I'd say. (They can probably get a 20% discount for volume, but that still only brings the price down to about $400K)
Agreed. Postgres has a proven reliability record, is free, is available on most platforms and is practically maintenance free. Oh and it scales really well on high core counts - without costing a penny more. I have used SQL Server in the past and it is good - but the alternatives are also good. And many of them have no licensing costs.
Small servers these days are shipping with 6-24 cores. Who knows how many cores they will have in 3 to 6 years, probably double that. Don't forget to build that into your budget.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
