Slashdot Mirror
Oracle Effectively Doubles Licence Fees To Run Its Stuff in AWS (theregister.co.uk)
Oracle has changed the way it charges users to run its software in Amazon Web Services, effectively doubling the cost along the way. From a report: Big Red's previous licensing regime recognised that AWS's virtual CPUs were a single thread of a core that runs two threads. Each virtual CPU therefore counted as half a core. That's changed: Oracle's new cloud licensing policy says an AWS vCPU is now treated as a full core if hyperthreading is not enabled. A user hiring two AWS vCPUS therefore needs to pay full freight for both, effectively doubling the number of Oracle licences required to run Big Red inside AWS. And therefore doubling the cost as well. The new policy also says: "When counting Oracle Processor license requirements in Authorized Cloud Environments, the Oracle Processor Core Factor Table is not applicable." That table says Xeons cores count as half a licence. Making the Table inapplicable to the cloud again doubles the licence count required.
Fuck that. He wants an island to moor it to.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
But if you're already planning on rewriting your software to work "in the cloud", migrating to a different database engine is not that much additional work.
It's nowhere near enough work to make their closed ecosystem an effective deterrent.
Continues onward.
What is the point in using Oracle software in ANYTHING these days outside of the support contract*
*Only reason I can think of.
Wheel of Time: Book by Book and Sumview (summary review) Bigdady92 style: http://bigdady92.blogspot.com/
If they want to price themselves out of business, so be it.
Just get a free copy of SQLITE, job done.
No sig today...
At this point you have to be mentally ill to develop anything new around Oracle and most of their business is legacy. They keep reporting double digit declines in new licenses.
http://blogs.wsj.com/cio/2017/01/17/oracle-co-ceo-mark-hurd-says-80-of-corporate-data-centers-gone-by-2025/
http://www.marketwatch.com/story/oracle-profit-software-license-growth-decline-2015-09-16
The surest way to kill legacy business is to make legacy more expensive than new build. Doubling cost will tend to do exactly that and thus speed up the decline. I don't get it. But then I'm a lowly code monkey and not a genius MBA so I'm probably missing something. (Sarcasm).
Anyway, I just want to live long enough to see them stop existing.
This move certainly won't slow down everyone's efforts to move off Oracle. Suddenly it's worth twice as many engineers to end the pain. Hopefully there aren't any Oracle victims left who haven't started on their "move off Oracle" plan - that would be sad, really.
Socialism: a lie told by totalitarians and believed by fools.
Step 1: Do this.
SQLite isn't remotely competitive with Oracle. It's nowhere near in the same league as even PostgreSQL or MySQL.
SQLite is a toy database with a huge amount of limitations that's found a niche in "I need a RDBMS for something simple, and rarely used". Thus the use for desktops to store things like configuration and music databases. In such cases it works well.
If you're even thinking at all of multicore performance, SQLite is not the database for you. It's got absolutely dreadful concurrency and will die under anything resembling a serious load.
Just get a free copy of SQLITE, job done.
Said no one who actually uses all the performance and scalabilty of an Oracle DB - EVER.
Not sure if you're being sarcastic or serious, because there are way too many naifs on /. that actually think you can do something like that.
Something like that'd be good enough for someone running a toy DB "in the cloud", but if you want real scalability and performance you're not running your DB on a VM on someone else's God-knows-what cheap commodity hardware whether it's Oracle, PostgreSQL, MongoDB, or whatever.
Time for the hadoop thunderclouds to rain on Oracle's revenue stream.
I think this is an excellent move for Oracle, and I enthusiastically applaud them. I think the company will experience significant revenue increases with this pricing change, and that's always a good thing. In fact, I encourage them to raise prices even more.
For all the naysayers, as I always say when someone complains about Windows, "if you don't like it, don't use it".
If hyperthreading is not enabled, then two "processors" by Intel's reckoning corresponds to two physical cores.
BTW Intel's marketers have pushed the idea that a hyperthreaded core is almost as good as two cores, but it's not, except maaaybe for special workloads of which almost never be the case for an enterprise relational database.
Said no one who actually uses all the performance and scalabilty of an Oracle DB - EVER.
Said the person who's using MySQL on Slashdot - This very INSTANT
That was the turning point of my life--I went from negative zero to positive zero.
Someone had to do it.
I can just see some exec racking their brain trying to figure out how to increase revenue finally had an epiphany: "Eureka! I got it! We'll just charge our customers twice as much!" to which everyone at the board meeting replied "Brilliant! You deserve a promotion!". Smiles and carefree laughter were gifted with abandon that day...
My eyes reflect the stars and a smile lights up my face.
Is there a technical reason for using Oracle over something else?
"Big Red's previous licensing regime [PDF] recognised that AWS's virtual CPUs were a single thread of a core that runs two threads. Each virtual CPU therefore counted as half a core."
That's flat out wrong; the actual old licensing regime counted each AWS vCPU as a full CPU core even though it was actually hyperthread. The new licensing regime counts each AWS vCPU as one half a CPU core (unless hyperthreading is disabled for the instance). That change alone effectively cuts the cost of licensing on AWS in half:
Old: Oracle running a instance with 4 vCPU (4 hyperthreads on 2 CPU): licensed as 4 CPUs requiring 2 cores of licensing
New: Oracle running a instance with 4 vCPU (4 hyperthreads on 2 CPU): licensed as 2 CPUs requiring 2 cores of licensing
The other change with the "Oracle Processor Core Factor Table" effectively doubles the cost back to where you started anyways:
"The intel core factor is 0.5, so an 8 core physical box requires 4 cores of licensing. Now on the cloud, an 8 core VM (16 vCPUs on AWS or 8 vCPUs on Azure) requires 8 cores of licensing."
This is the gospel of our economic system, right?
“He’s not deformed, he’s just drunk!”
Said no one who actually uses all the performance and scalabilty of an Oracle DB - EVER.
Said the person who's using MySQL on Slashdot - This very INSTANT
SQLITE is not in the same league as Oracle or Mysql. SQLITE is basically one step up from a flat file.
Oracle: Trying to price ourselves out of business since...well...SINCE!
Chas - The one, the only.
THANK GOD!!!
It seems like this move could mean one of two things: either Team Oracle thinks that there is sufficient willingness to pay among users of their products, and they were previously leaving money on the table in AWS instances; or they fully expect this to seriously dent use of their products in AWS; but don't care because they have their own 'cloud' offerings and want everyone not running on premises to be buying cloud from them.
Any guesses as to which it is? Is this a "Larry's a jerkass; but he knows that most of us will suck it up and pay the extra" situation; or is this a straightforward move to make one of the more popular cloud options blatantly uneconomic for use with Oracle stuff, in order to improve the apparent value of Oracle's pet cloud?
SQLite is fopen with searching built in.
Don't get me wrong; it's amazing at being what it is. But it is basically a fancy fopen.
Setting aside the likelihood that you were tripped up by the minor ambiguity between "an Oracle DB" and "a DB other than the Oracle DB but still owned by Oracle Corporation", using a website is very different from using a database, even if the website happens to be using that database. That's a double-equivocation building up to nothing more than a trivial nitpick.
Depending on how you're deploying it putting your database system in the AWS cloud doesn't require a software rewrite at all - the whole thing can be transparent to the application (much like just using a virtual server over a physical one already was).
Changing the database from Oracle to something else isn't quite so simple.
"People who think they know everything are very annoying to those of us who do."-Mark Twain
Although I agree with your general assessment, and I think the grand parent was just joking... I want to clarify somethings about SQLite before your post misinforms some of the visitors to this site.
SQLite is the most deployed database in the world. Oracle has more of a niche use case than SQLite. The functionality footprint is extremely small. The installation library is smaller than many DB connection drivers! In short, it provides SQL syntax based access to a flat-file in RAM or HD. It is simple and neat; yet provides ACID compliance. Programming environments and languages do not provide SQLite connectivity; they incorporate the entire system as a library.
It is the storage & decision mechanism for many mobile applications. It is utilized in many embedded & SOC systems. Although the library itself is single threaded, it supports concurrent access. So you can actually write programs to be multithreaded/multiprocessed to scale with the number of cores/CPUs. I personally have written programs that trade CPU counts & RAM for execution time.
But anyway, the cross section of use cases for Oracle/Postgresql/MSSQL and SQLite are basically non-existent... maybe you see some overlap in Prototyping to Deployment. MySQL and SQLite do appear to have some minor overlap, but its small there too.
and Oracle is evil. These are all self-evident. For big companies, I guess the pain of migrating to another solution is greater than the pain of bending over for ol' Larry. Especially since they just pass the additional costs along. It's a pity, though, that some lesser evil doesn't take on the task of de-throning Oracle - it would be sweet to see them losing market share fast, and even sweeter to hear them begging their lost customers to come back.
'The Economy' is a giant Ponzi scheme whose most pitiable suckers are the youngest among us and the yet-unborn.
When is the last time someone failed an oracle audit? I think they are similar to red traffic lights in LA, mostly suggestory in nature.
We can have Embedded Firebird these days anyway.
Ezekiel 23:20
Looks like Oracle was not checking this flag and was counting it as half a cpu even hyperthreading is off. They corrected this. People who turn on hyperthreading will see their virtual processor be counted as half a cpu.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
When hyperthreading is enabled, it is 2 vcpus per core.
When disabled it's 1 vcpu per core.
Licensing per core should follow per core, regardless of vcpu count.
What's there to complain about except that some folks that were taking advantage of the licensing run-around got caught? Big whoop.
It is hard to feel sorry for anyone that gets into bed with Oracle. They deserve all the fleas and STD's that Oracle gives them.
You can't even call Oracle's customers victims since they gave Oracle permission to do this.
Oracle and SQLServer are for people who like large, sharp objects in their ass.
MySQL/MariaDB is for people stupid enough to not know Postgresql exists.
SQlite exists for the use cases where Postgresql is overkill.
have to deal with their licensing! Our last Microsoft audit cost us over $50k. It took us weeks of time between accounting and operations. We don't know how much the Oracle audit cost since we lost count since it was so much more of a hassle. They required us to run a VM on VirtualBox on our production network with Ethernet bridging enabled. It shutdown our network since we don't have STP (spanning tree protocol) enabled. We lost two customers and several tens of thousands of dollars of monthly income because of that.
While I agree that SQLITE isn't nearly the performance or features of Oracle, it's amazing how many places using Oracle because the name and cost impress the execs on the golf course could actually do just fine with PostgreSQL or Mysql.
It is not simple if you were stupid enough to use proprietary extensions to SQL.
That has a different pay structure versus trying to bring your own license and pay more.
SQLite is the most deployed database in the world.
And Fox News is the most popular "News" source.
I use SQLite for all sorts of things. As a personal database, it's great.
But it's not designed for industrial use. For that we have more powerful open-source DBMS's such as PostgreSQL or MariaDB.
When a shrinking company suddenly jacks up their prices, it usually means they are milking the short-term for all they can and don't expect to be around in the long term (at least not as a real company).
Big tech co's basically have 4 choices when they start slipping:
1. Innovate and keep up
2. Hold existing customers hostage and milk them dry before they finish migrating away
3. Sue other co's using sketchy patent claims, hoping for at least settlements
4. Wither until you are bought out by a holding co. that does nothing useful
(Not necessarily mutually exclusive.)
Table-ized A.I.
Not just execs. Also Mysql is Oracle's entry level RDBMS. I hate Larry as the next guy but you need to give it to ora is a very good product.
*chuckle*...
Oracle is good for only one thing these days: RAC.
For everything else, there's PostgreSQL, MSSQL (if you're into that), MySQL... and funny enough, a simple heartbeat script with STONITH capability, some decent replication tweaks, and a load balancer in front of the cluster can do most of what RAC can do on a practical level - at least enough to not really justify the $$$$$$$$ spent on RAC (again, in most cases.)
Put this way: Unless you're, say, running a brokerage and use a RTOS for sub-millisecond trading in front of that DB, or have a multi-zillion-dollar legacy system that was built around Oracle proprietary SQL/code? Hate to tell you, but Oracle is not really a standout these days.
(well, if you're an Oracle-certified DBA or analyst, I guess you can contrive an excuse or two to justify spending all that certification money...)
Quo usque tandem abutere, Nimbus, patientia nostra?
I used to use Oracle (the DB) for years (versions 5, 6, 7, 8, 9). Then I discovered it is not necessary to be shafted!
Royalties? hell no, we won't pay - no f'n way! (Sure there is a 70's soul track to that effect somewhere).
Yep. FreeSWITCH uses SQLite as its default database and is fine for low to moderate loads. Anything serious requires something like Postgres, which it also handles.
slashdot: A failed experiment.
Not often you hear "Oracle" and "effectively" in the same sentence!
Sent from my ASR33 using ASCII
Oracle, Microsoft, and Adobe Systems are competing to see who can be most abusive.
How can you compete with the biggest cloud vendor? Make it look more expensive.
It's unsurprising how many place use Oracle because a key software package only supports Oracle or MS SQL databases.
SQLite is the most deployed database in the world.
So? How often do those /databases/ need to be anything more than a JSON or XML doc? How often is said SQLite DB a technologically advanced, elegant, or better solution?
/can/ create a database doesn't mean someone /needs/ a database.
Just because someone
Thirty four characters live here.
Oracle is shite.
Saying sqlite is the most deployed database is like saying thermostats are the most depolyed computer in the world. Might be technically true, but really...
So now each thread on a hyperthreaded core counts as one? No wonder why AWS virtual cores suck balls.
MongoDB grabbing lots of Oracle customers
Wait, what? Nice try. You almost had us.
Breakfast served all day!
Unless you are doing some pretty hairy things with your DB, there's really no reason not to move to Postgres for heavy lifting, or other DBs for more trivial workloads.
I can't imagine running Oracle in a Cloud would be a good idea in the first place. Then again, I don't know everything - perhaps there's a valid use case to do so. But really, if your DB is important, I can't see virtulizing it in anything other than a VM meant for only DBs, not a general cloud VM. EG: OK for something like DBaaS, but not to fire up a VM and just go like it's dedicated hardware.
Necessity is the plea for every infringement of human freedom. It is the argument of tyrants; it is the creed of slaves.
STONITH
RAC
Next time, try the English language.
Oracle profit, software license growth decline (Sept 16, 2015)
Very funny.
http://m.slashdot.org/story/31... "MongoDB CEO Claims They're Luring Customers From Oracle"
AFAIK Postgres has a PL/SQL compatibility module now.
In most of these cases, SQLite is a better solution than JSON or XML.
JSONs negative is that it is basically a direct translation of the internal data structures of a program. This results in an inflexible design that is difficult to reuse across programs and versions. SQL provides a universally recognized, mature, stable interface that is tried & tested. It converts the data structures to and from a well organized standard. SQLite provides that capability at an extremely small footprint.
XML... is horrible. In the simplest cases, you should use a .conf file. In more complex cases it is no more useful than SQLite. However, it comes with a massive library footprint. And if you need any kinds of transformative views of your data set, it comes with a large overhead. Additionally, both of the above sit in memory while SQLite doesn't have to.
But to address the heart of your post.. what's your point? I have seen many Oracle, PostgreSQL, MySQL, MSSQL, etc instances that were clearly overkill and/or misused. Situations of _must_ use it because you have it and Everything is a nail. We still count these as DB deployments. In fact this true of most IT tools.
SQLite is a single user database that stores a lot of data for a single user quickly efficiently and reliably. To say that there is no industrial use case for it is to say that there is no situation where a single user application needs to store a lot of data. Furthermore it would be overkill to use a database server to store some application data.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
It was designed for military use.
Most amazon instances have hyperthreading enabled, so 4x increase is for the most part avoided. Oracle SE2 is licensed by number of sockets, so no change.
However, Oracle EE with Bring yuor Own license would suffer from the removal of core factor .
Folks should just move to postgresql or enterprise DB instead.
If you're not using Oracle's extensions to wring out performance from your RDMS, you're probably wasting even more money than you were by choosing Oracle in the first place. No one chooses Oracle to get MySQL or PostgreSQL performance.
Don't know about MongoDB, but the cloud part is surely a difficult one to crack. There are really just 1 cloud vendor and that is amazon, and a couple of distant also rans as google and Microsoft. IBM are building up, but if you're not top 3 might as well not play, same goes for Oracle.
As for java licensing, not sure how that would work, so far it was free unless you used it for embedded systems.
Bread and butter for Oracle will still be their enterprise software stack, but with amazon offering more and more complete end to end solutions it is easy to see the writing on the wall that Oracle is in for a tough ride.
The first wave of migrating to cheaper x86 public and private cloud infrastructure is happening, and some new huge successful companies have gotten off the ground at huge scale using other than Oracle software stacks. And increasingly, large enterprise customers are trying to refresh their legacy stacks and have IT that looks more like netfix or uber.
", running a brokerage and use a RTOS for sub-millisecond trading in front of that DB"
I'd think that would be a very bad idea. Oracle has a lot of latency, overhead and unpredictability that would make eliable sub milllisecond response times difficult to achieve. (redo writing, archive log copy checkpoints, stats gathering, shared pool filling up, etc. can all cause multi ms spikes ) Pretty much any other DB would be more suitable out of the box. Oracle would need to be heavily tweaked and tuned to be a contender, and even then the gotchas are piling up.
There is a reason for this. Oracle has a Cloud service! So, of course they will increase the license fee for other cloud services that use his products!
And PL/SQL is supported on SQLite through a commercial product StepSqlite (http://www.metatranz.com/) which is in limited alpha now.
Interestingly, Oracle's own BerkeleyDB also uses StepSqlite for PL/SQL (http://www.oracle.com/technetwork/database/berkeleydb/db-faq-095848.html#DoesBerkeleyDBsupportPLSQL).
SQLite has PL/SQL support as well, through a commercial product StepSqlite (http://www.metatranz.com/) -- in limited alpha now.
Interestingly, Oracle's own BerkeleyDB also uses StepSqlite for PL/SQL support (http://www.oracle.com/technetwork/database/berkeleydb/db-faq-095848.html#DoesBerkeleyDBsupportPLSQL).
Wasn't this the reason that AWS came up with "dedicated hosts"?
https://aws.amazon.com/ec2/ded...
...You are over-qualified and under-paid. If we give you a raise, we will break the cosmic balance of the universe.
The license may prevent you from benchmarking Oracle's database, but wouldn't a third-party guest be able to do so not being a party to Oracle's licensing agreement, especially if they did so without the license consentee's knowledge?
Given how desperate Oracle looks lately, how much do you think they'd sell Java to Google for in an effort to prop up their core database business?
Like like Red said in Shawshank Redemption some times people get Institutionalized...
It isn't so much a technical issue, but one of momentum, which is hard to change in anything but over a long time. Like MS, Oracle has been around and dominate for a very long time. Which means their install base is large. Which means a lot of things use it. Which means a lot of people are trained in it and familiar with it.
The difficulty is that IT shops have to maintain and support applications and databases. having to support more than one is costly. porting existing ones is costly. Training or hiring new staff is costly, etc...
So if you are a new organization, then sure you likely have the flexibility to easily review and select whatever fits your needs. However if you are an existing organization, which most are, and most of your infrastructure is already Oracle, well that is a pretty hard decision for an IT head to try and reverse and change.
However licencing things like this do take a toll, and you will find that organizations start to try to experiment supporting more than one, and expand upon it (we've been dabbling with MS SQL for a number of years now). It is easier to propose to executive management the additional expense and overhead when you can show the increase in existing licencing costs. Over time if Oracle isn't careful will find itself in trouble. However that is likely not anytime soon, so for now, the bean counters just count more beans.
Oracle is dog slow. All that crap added to protect data slows things down.
Oracle is no faster than PostgreSQL, in many cases it is significantly slower.
So you are right, no one chooses Oracle to go slower, they do it for other reasons.
Extreme idiocy and lemming behavior is the primary reason.