"...the only database in the planet..."
Oracle is industry’s single largest database vendor – which was great during the days before cloud and open source. Now, however, the on-prem RDBMS giant faces a challenge, one which is compounded by cloud. According to Gartner, Oracle owns just a tiny sliver of the cloud infrastructure market – 0.3 per cent. But while that …
There is one valid point in the whole thingymagig - Oracle installs are not at all frictionless. The rest is a fair bit of hyperbolery, esp Mongo stuff.
On macOS or Linux you can't just install SQLNet connectivity. You need to download it, accept all sorts of useless download conditions and only then can you install it. Ever heard of apt-get, macport or homebrew? Guess not. Looking to auto install it with Chef, it was a mess because they only provide it in RPM format, not great on Ubuntu. Lots of funky and unstable looking Chef packages to deal with that - I ended up just doing an apt-get install alien an inelegant but simple hack.
Not the database itself, which, of course, is costly. Just the connectivity driver, which is free. Which of course sucks to configure - tsnsnames.ora, anyone? And sqlplus deserves to be roasted, quartered and staked out on an anthill, in whichever order makes it the most painful.
Matt does have a point that complicated and expensive licenses really get in the way for starting up. This particular issue is, I believe, also going to bite MS hard, which is why they are pushing Linux hard - they know Linux is easier to experiment with, unless they bite the bullet and free up Windows licenses outside the enterprise crowd. Lose enough experimenters and you may start losing real business too.
If your needs can be met with Postgresql, it is a brilliant little piece of RDBMS tech. I can definitely see it eating up smaller Oracle db business for custom apps - where the db is not already predetermined by a vendor - going forward. I'd love to know how well it scales up in the GB ranges, but it really is impressive. Very much unlike MySQL, which isn't.
Still, at the end of the day, when you really need a big big RDBMS and have $$$$$$$$$$$, Oracle is not the worst choice. Neither is SQL Server.
Still, at the end of the day, when you really need a big big RDBMS and have $$$$$$$$$$$, Oracle is not the worst choice. Neither is SQL Server.
Wgen speaking of Oracle, I think you need more dollar signs there. Remember, five dollar signs *per core* (whether the cores are actually used for the DB or not, such as charging for ALL cores in a VM hosting system, even if Oracle is only running in one VM).
Already fixed, but thanks for pointing it out.
Matt (Asay) has articles on the Reg going back to at least 2012; in that time it looks like he has had about a million jobs (Sorry, Matt!) and was, for a short time, employed by Mongo/10Gen to run their community. He is now working for Adobe and there are stints for Alfresco, Canonical and others mixed in there somehow.
I think he highlights a good point regarding the revenue share of OSS DBs having a magnifier effect on the loss of revenue of Oracle. The NHS switched from Oracle on the Spine, for example, saving a whole bunch of money. Shame they went to Riak, but it’s easy to judge with 20/20 hindsight.
In June 2012, he wrote this puff piece on MongoDB. 3 months later he was hired by 10gen (now called MongoDB Inc), where he had "several areas of responsibility while at MongoDB, including corporate strategy, business development and corporate marketing" and "Added marketing leadership in December 2013, covering web, corporate marketing, community and communications. Shifted business development in May 2014 so as to focus full-time on marketing. Helped to cement MongoDB as the second-most discussed (and top-4 most popular) database"
Make your own minds up whether this article has any bias in it.
Many publications have an editorial policy that where an author has (or has had) an interest that might potentially influence the perspective of their article, they are required to note this in the article footer.
If Matt Asay has recently worked for MongoDB then that should have been noted. (I am not suggesting that that necessarily skews his opinion (after all, he is apparently no longer working there), but I do think this should have been noted, so that readers can decide whether or not that might reflect any slight bias in the article (again, I am emphatically not suggesting this, but being open and transparent about these things is better).)
"As it says at the bottom of the article, Matt is head of ecosystem at Adobe. He left Mongo DB in 2014."
But still pushing cloud, e.g. "a developer's first decision is what cloud platform they'll use". My first decision would be "Does it matter if my data ends up on haveibeenpwned?" and take choice of storage place from there.
"But open-source databases are generally outperformed by MS SQL and Oracle on benchmarks on the same hardware, so you would have higher hardware costs for the same performance?"
We ported an in-house business app from Oracle to Postgres to save on the horrendous licensing costs. An unexpected outcome was that the app ran between 3-25% faster on the same kit after porting. That was almost ten years ago and it has run without a single hiccup since then.
I can't vouch for the performance of MS SQL as it wasn't even considered as an option in this case. Like Oracle, MS SQL must be licensed for every physical CPU core in the underlying hardware, even if they are not presented to the DB server. The alternate option is to buy a CAL for every user of the app connected to the database it serves but that's also insane. Either way resulted in very little practical cost saving over Oracle, so even if MS SQL turned out to be significantly faster than Postgres (rather unlikely) the cost difference would still not have justified it.
What a dire article.
"the primary choice is between PostgreSQL and MongoDB"
No it isn't. Really it isn't. On AWS it's Aurora, Oracle, Postgres, SQL Server, Dynamo, Redshift, Athena or MySQL. Probably more I forget. And then there's all the prepackaged servers in the AMI library.
"MongoDB, gets picked when a developer is refactoring her application and needs a significant boost in developer productivity "
Only if your developer doesn't understand data models...
"The database they elect to use then follows from the options made available by that cloud platform"
Nope. I run BigQuery on GCS, Redshift, Dynamo, VoltDB and Aurora on AWS and Azure Analysis Services. Use the right tool for the job.
I wish we could downvote articles, because this is complete rubbish. That Oracle has lost in cloud is not news, everyone knows it. And doesn’t the choice of a cloud platform necessarily include first assessing what technologies it makes available? Who chooses a cloud first then decides later what they want to use? Only idiots. Or MongoDB users.
"Who chooses a cloud first then decides later what they want to use? Only idiots. Or MongoDB users"
Or the people paying you money to develop things and have heard about this "cloud" thing and want in. (There are a lot of these)..
I dont know *anyone* who would consider using Oracle for a new project. Even if you can get past the licencing aspect, that Stupid installer needs to be taken out the back and shot, along with the intern who wrote it. (It is ridiculously difficult to install on a headless server, and you have to install unnesseery stuff you will never use again like all the X libraries, and I think its really about Oracle figured out "Program files" on Windows Clients and servers. Its been over 20 years now)
Administration is a pain in the arse, unlike every other modern database. (I remember DB2 comming close, but that was a long time ago)
Its possible at the very highest end, that Oracle may out-perform other databases, but how many people really ever use the full potential of an Oracle installation? Or even a MySQL installation for that matter?
Oracle will not go quietly into the night, what this peace misses is the absurd level of lockin a mature ERP system has it is a total nightmare costing megabucks for any organisation regardless of their scale.
Oracle has a long history of excelling in these back-end markets (Even if I want to hang all of their
auditing sales people).
Yes, you can get away from the $$$$ oracle licences and go cloud but if the migration costs 10+ years of licencing as well as running the risk of killing the company in the meantime people won't move.
All told Oracle can coast in the wrong direction for a long time before the shareholders will kick off and insist on change.
> Oracle will not go quietly into the night,
True, and don't forget the huge government lock-in it has "achieved". But the article is valid in spite of the criticism of some commentards, as Oracle failed to understand its Sun acquisition for MySQL, and certainly failed to understand the Free and Open Source path that My SQL offered.
But there will be many El Reg readers whose livelihoods depend on Oracle, and their jobs are almost certainly safe for years.
True, and don't forget the huge government lock-in it has "achieved".
Some of that "lock in" which appeals to Government is down to Oracle being willing to pay for security certifications that open source projects can rarely afford or achieve. It's hard to get freely-available software certified when anyone can download it, change it, build it with any of a dozen compilers, etc.
There's also the performance issue. Mongo & pals are fine for small DBs, but for the really huge DBs that governments and big corporations use there are really very few RDBMS that can cope.
The bureaucracy inherent in acquiring Oracle's database is almost nothing. You can download any version you wanted from their website(checked again now just to be sure - http://www.oracle.com/technetwork/database/enterprise-edition/downloads/index.html) without anything.
They purposely make it easy to lure people into using it so then they can come back with audits.
Last time I seriously dealt with Oracle DB licensing was about 10 years ago, and at the time it was pretty easy. As a new hire I tried to advice the company how to deploy Oracle properly when they were undergoing an audit. My manager decided to ignore me, they paid their fines to get back to a normal stance and kept going. Until the 2nd audit came around when I once again told them what they needed to do, and they did it(I guess I did it as I did most of the work) that time, still had to pay fines but they were legit fines they were terrible/lazy about managing their licenses. I found it ironic I knew more about Oracle licensing than the Oracle reps did at the time (specifically around leveraging Oracle standard edition on multi core processors). I also did things like install single socket(quad core) vmware hosts(which vmware did not "support" at the time), to get more Oracle instances up (even though Oracle did not "support" vmware, I think they still may not). For production it was all bare metal and optimized with fast dual core cpus or quad core depending on EE or SE licensing.
That particular company when they started had Oracle SE "One", the tiny DB. Then they hired an Oracle DB consulting firm to help them manage the systems(this was before I was hired), the first thing the consulting firm did was to install Oracle EE everywhere. Company was hit hard for that filing support cases against EE when they were not licensed for it. Later on the 2nd audit got hit again because that DB consulting firm had monitoring that used partitions in oracle, another expensive add on. No other apps or anything was using partitioning but company got hit with the bill for using an unlicensed feature. The monitoring software was then updated to not require partitioning.
Previous to that company was a place that had massive abuse of Oracle licensing we had probably a dozen or more hosts, and were only paying for a couple(everything on EE). For some reason either Oracle didn't bother to audit, or when they did audit we got by somehow (I wasn't responsible for those systems). Eventually the company got correct in their licensing but took a few years.
I think Microsoft is similar, though at least with MS you can't(as far as I know) download their biggest products for free and be able to use them without a license key/file etc.
I have used Oracle DB for the past several years just as a back end database for VMware vCenter, very low utilization. Plan to move to vSphere 6 this year and to the vCenter appliance clustering along with it (Postgres I guess), so won't have Oracle anymore after that. They ping me every so often to try to get more sales but that doesn't go anywhere, and they haven't expressed any interest in an audit(for what I licensed I know I am way over licensed vs what is actually used). Oracle actually sent me an email recently reminding me my support is expiring on Feb 8 2020 (so why email now??) and the renewal fee is $3.15 (no idea where that number came from). They have been emailing me for a year saying my support is expiring in 2020 and I should renew. I mean I can understand emailing a few months before expiration but more than a year? Never seen that before.
It really would be nice if Oracle Enterprise Manager's features were available in Standard edition. I loved OEM at least the performance management stuff being able to see what queries are doing. I recall 10 years ago again it was possible (not "legally") to enable those features in Standard edition, then when the audit came I could just wipe out the data stores for OEM and replace them with regular ones, then reverse it again later(didn't care about data retention). Though with 11G that doesn't seem to be possible anymore(at least not in the same way it was then). My Oracle is really rusty these days though(I have never been a DBA).
For a professional DBA, "manage" in this context means "envisage the horror of managinig", quickly followed by "rapidly find a new job where their boss is not a gibbering fuckwit who would even suggest that managing a million databases was some sort of actual real practical possibility".
How is Mongo a migration target from Oracle? It's not a relational database...
PostgreSQL - absolutely. That is where the majority of our Oracle databases are migrating to. And with the PostGIS extensions, it also replaces the Oracle Spatial option as well for the location part of your data.
The only other database that is of any interest that isn't open source based is NuoDB, due to it's unique scaling abilities especially for use with containers.
Of course, MS Sql Server has the same problems as Oracle - old monolithic technology at a high cost. Why anyone would build anything new with either SS or Oracle is a mystery.
I was about to chime in with a comment that most people wouldn't choose Oracle not because of their technology, but because of their money gouging corporate policies.
Then I read the bottom line of the article and realised that the author lives in a glass house when it comes to money grabbing corporations:
Matt Asay is Head of Developer Ecosystem at Adobe.
Yeah, I can see why you didn't mention Oracle's evilness much...
"It would be comforting to the Oracle faithful to believe that this open source onslaught isn’t having an impact on the database behemoth."
No, it will be comforting in the competition drives down the price.
People using Oracle typically are leveraging the Oracle technology, and switching en masse to an open source "alternative" is a non-starter.
Smaller shops that just want to hold data in tabular form and get at it with SQL are a different use case than the large enterprise, where replication technologies and resilience are paramount.
I know if the license calculus were to change favorably that my director-level people would be a lot happier.n But telling them to move to Mongo DB or MySQL would get me fired, free or not.
At least the official version? Sure MariaDB seems to be the more popular variant of that DB, but it's not as if you can't get MySQL directly from Oracle.
If customers don't want to pay for support it doesn't really matter what you make. The company I am with used to pay for Percona MySQL support(we use Percona across almost everything though there is a push to go to Maria). Percona's pricing was very attractive at one time I think it was basically $15k/year for unlimited support unlimited instances. We filed maybe a few cases a year, not much at all. Then one year it jumped to something like $120k/year for the same stuff(actually think it was less but not certain this was 2-3 years ago) so it was decided to drop their support and stick to internal staff only. Today I don't see pricing on Percona's site so not sure what it is now.
We tried RDS years ago when we first launched the app stack, it was just terrible. I'm sure it is probably fine for pretty generic setups, but the lack of control was just maddening. Getting data out of the thing was quite a mess too, when we finally moved out of amazon cloud in 2012 we had to do mysqldump to get the data out to import to real mysql servers, a process that wasn't quick.
Last I recall Amazon themselves were huge users of Oracle internally having a site license(s) -- licenses that made it cheap/easy to deploy everywhere. I have a friend who has been an Oracle DBA manager at amazon for 12 years now, haven't talked to him in a few years though.
22 years of working with Oracle and the writing is on the wall. Sure there's not going to be huge uprising and Oracle will not end up like Filemaker, a niche little vendor but Oracle software is now so fat and bloated, it's basically the MS Word of the RDBMS world.
Practially every company I've worked for in the last 5 years is actively testing alternatives for RDBMS tech, with Oracle falling behind even SQL Server for consideration! With the last few shops looking at AWS and seeing that you simply pay for the storage and usage of MySQL and PostgreSQL, not to mention toying with NoSQL ( I'm not a huge fan myself ), the world of databases is changing and developers are now finding that we DBAs can be side stepped! Ha ha! (*)
( * I'm not saying I like it or even that it's a good idea, but the two camps have been at war for many years and as a battle hardened DBA I can see that we're slowly losing the fight to the developers, the devops and the architects. )
Oracle is still the best choice for the largest and most critical databases. The problem is that the alternatives are a lot better than they used to be. Most organisations dont need the scale and performance provided by Oracle with many alternatives that are "good enough" Also there is a blurring of the lines between the NoSQL and SQL databases. There are no "pure" databases anymore with many of the best features of all the different approaches (e.g. NoSQL features added to Oracle and ACID and resiliency features added to many NoSQL databases) going to producing better databases overall.
IMHO this discussion will become academic because Serverless computing e.g. AWS Lambda and SaaS will change who chooses the database any from the customer.
Much of the points are good, although there is a lack of real technical understanding of why DB platforms are used.
The Open source RDBMS are limited by both non-functional (think security and HA/DR) and functional features (think temporal tables, functions, developer options). However for basic websites or "DIGITAL" as it is known now rather than a complex transactional application Open source is often a good fit.
I don't think migration from RDBMS to a document, graph or other type of DB (e.g. MongoDB) is one that is usually taken, maybe for some functionality but this is a major re-platform of the application and application codebase (think in terms of millions of ££ of dev effrot)
Adobe talking about ripping off customers that's rich!
"PostgreSQL, in particular, has made it relatively straightforward to migrate things like stored procedures from Oracle to PostgreSQL."
Wait, wait ...! PostgreSQL doesn't even *have* stored procedures! (But it will in the upcoming version 11.) I think the author actually means MariaDB which has implemented a subset of PL/SQL in its most recent version, 10.3. MariaDB now also supports a lot of other features found in Oracle, such as sequences.
Full details here:
Also: "Widen that aperture a bit and you get MariaDB, a fork of MySQL, in the 15th spot."
That should read "13th spot"! Check it out:
Just saying :)
RDBMS? Really? How out of touch do we have to be to make this distinction between relational and flat database? Have we not been in relational database land for oh, I don't know, 30 years or so? Sheesh. Call it something else. Get in front of it or become obsolete. RDBMS? I had to stop and remind myself what the R stands for. Sorry. I'm not a programmer, but still.
Analysis Under Nevada's baking summer sunshine, Snowflake last week promised it would bring together two ways of working with data that mix about as well as oil and water.
The data warehouse vendor – well known for its stratospheric $120 billion post-IPO valuation – said it would support both analytics and transactional workloads in the same system.
Launched at the Snowflake Summit 2022 in Vegas, Unistore would be the "foundation for another wave of innovation in the Snowflake Data Cloud," said Christian Kleinerman, senior vice president of product. "Similar to how we redefined data lakes and data warehouses for our customers, Unistore is ushering in a renaissance of building and deploying a new generation of applications in the Data Cloud," he said.
YugabyteDB, the self-styled double-decker distributed relational database, has introduced a read-committed isolation level, allowing for more flexibility for devs and bringing it into step with its more established RDBMS rivals.
With the general availability of YugabyteDB 2.15, the company behind the PostgreSQL-compatible database, said to be inspired by Google Spanner, claims to be adding performance enhancements and dynamic workload optimization.
Speaking to The Register, CTO Karthik Ranganathan, a former Facebook technical lead, said Yugabyte was the first and only distributed SQL vendor to offer all three PostgreSQL isolation levels, making it that much easier to shift existing applications to a distributed database with advanced support for the default PostgreSQL isolation level.
DataStax, the database company built around open-source wide-column Apache Cassandra, has launched a streaming platform as a service with backwards compatibility for messaging standards JMS, MQ, and Kafka.
The fully managed messaging and event streaming service, based on open-source Apache Pulsar, is a streaming technology built for the requirements of high-scale, real-time applications.
But DataStax wanted to help customers get data from their existing messaging platforms, as well as those who migrate to Pulsar, said Chris Latimer, vice president of product management.
Analysis At MongoDB's recent conference in New York, the company demonstrated its ambition in taking on workloads from other databases.
The company has made significant inroads into the database market with a developer-friendly distributed document database to help devs build modern, web-based, transactional systems.
Time series and search have become targets, with the promise of support for secondary indexes in the former, and Search Facets to help developers build search experiences more rapidly in the latter.
DataStax, the database company based on the open-source Cassandra system, has secured $115 million in funding for a $1.6 billion valuation.
Led by the Growth Equity business within Goldman Sachs and backed by RCM Private Markets and EDB Investments, the latest round follows a strong first quarter based on the popularity of DataStax's Cassandra DBaaS Astra DB. Existing investors include Crosslink Capital, Meritech Capital Partners, OnePrime Capital, and others.
Cassandra is a distributed, wide-column store database suited to real-time use cases such as e-commerce and inventory management, personalization and recommendations, Internet of Things-related applications, and fraud detection. It is freely available on the Apache Version 2 license, although DataStax offers managed service Astra on a subscription model.
PostgreSQL co-creator and MIT computer science professor Michael Stonebraker has listed his top requests for features to add into the popular open-source database, including a time travel function he admits was implemented badly in the 1980s.
Speaking at Postgres Vision conference, Stonebreaker said the time-travel code was fundamentally a good idea, as it allows users to query data from the database’s history, and had introduced an implementation in the 1990s.
“The problem with time travel back in 1995 was that my implementation was absolutely awful. It was slow, slow, slow, slow. Whether or not time travel was a good idea, the implementation was too slow to be interesting, so it was rightfully deleted by the powers that be in the committee that has been dealing with Postgres since 1995,” he told attendees.
Oracle is planning to build a national database of individuals' health records for the whole United States following its $28.3 billion acquisition of electronic health records specialist Cerner.
In a presentation, CTO and founder Larry Ellison said electronic health records for individual patients were stored by hospitals and physicians, and not replicated or shared between providers.
"We're going to solve this problem by putting a unified national health records database on top of all of these thousands of separate hospital databases," Ellison said.
MongoDB, the company behind the document store database, has unveiled columnstore indexing designed to help developers build analytical queries into their applications.
Set to preview later this year, the feature is designed to allow developers to create a purpose-built index to accelerate analytical queries without requiring any changes to the document structure or having to move data to another system.
MongoDB chief product officer Sahir Azam told The Register the feature would be available in the database and Atlas DBaaS to support human-like decision making inside the application based on live data.
MySQL pioneer Peter Zaitsev, an early employee of MySQL AB under the original open source database author Michael "Monty" Widenius, once found it easy to identify the enemy.
"In the early days of MySQL AB, we were there to get Oracle's ass. Our CEO Mårten Mickos was always telling us how we were going to get out there and replace all those Oracle database installations," Zaitsev told The Register.
Speaking at Percona Live, the open source database event hosted by the services company Zaitsev founded in 2006 and runs as chief exec, he said that situation had changed since Oracle ended up owning MySQL in 2010. This was as a consequence of its acquisition that year of Sun Microsystems, which had bought MySQL AB just two years earlier.
Cockroach Labs has finally added a new command line tool with the release of version 22.1 of its eponymous database, out today.
Although it was possible to deploy CockroachDB using something like Terraform (for example, deployment on Oracle Cloud Infrastructure) the process is often not particularly elegant.
"Until this release we didn't have an API to control the database," Jim Walker, recovering developer and product evangelist at Cockroach Labs told The Register during 2022's EU Kubecon in Valencia, Spain.
Biting the hand that feeds IT © 1998–2022