Only 5 weeks after completing the acquisition of Virtual Iron, a promising virtualization start-up, Oracle goes out and kills it. To make matters worse, Oracle has done it in a very brutal way. In 11 days even existing Virtual Iron customers will not be able to buy additional licenses. Or to quote the The Register article that broke the news:
“So basically, anyone that built their hosting infrastructure on VI…is now totally in the shit,” that partner tells us. “Unless they buy a whole bunch of licenses before the end of June, they will be unable to buy any more node capacity for their clusters. Oracle are shutting down the product, without giving customers some sort of replacement. That’s a huge customer/partner channel shafting.”
So, if you are a poor chap who convinced your CIO that virtualization is the right way to go and that a start up called Virtual Iron is the right way to do it, you may want to consider freshening up your resume compliments of Oracle. The worst part is that you were probably right to choose Virtual Iron as it indeed had some very good technology.
I’ve blogged about my opinions about what the future holds for MySQL customers in light of Oracle’s acquisition of SUN and MySQL. I predicted, contrary to the opinions of some of the industry pundits, that MySQL within Oracle will wither and die. I stand behind my prediction; in my mind it is not if but when MySQL will cease to be an Oracle offering. A number of people disagreed with me, going as far as accusing me of spreading FUD (Fear Uncertainty and Doubt). Indeed, there is a lot of fear, uncertainty and a lot of doubt in the MySQL customer base; it is not because of my little blog spreading FUD. I am guessing that this precedent with Virtual Iron is going to spook a lot more of the MySQL customers.
My advice for anyone paying for MySQL Enterprise is start calling alternative providers like Percona.com, or Monty Program AB. At least these folks really know what they are doing with MySQL. Most of the key people have already left MySQL even before Oracle announced that it was buying SUN and hardly any are expected to stick around after the acquisition is complete. True, both Percona and Monty Program are very small and are not likely to be able to take on majority of the MySQL customers but what they lack in size they compensate for in their expertise. For the lucky few who go with them before Oracle pulls the plug on MySQL Enterprise I think it will make a big difference.
There are other options. You can decide that you really don’t need support (not likely for enterprise users) and just go with one of the MySQL open source forks and rely on the community for help. Or, you can contemplate switching to a different database. PostgreSQL is a good alternative but if you are looking for support you would again be relying on a small player like EnterpriseDB. I think that for anyone using MySQL Enterprise i.e. those relying on vendor support, DB2 Express-C makes a lot more sense. Just like MySQL, DB2 Express-C is available for free to use in development and production environments. Unlike MySQL, it can also be redistributed free of charge by ISVs as part of their solution. And just like with MySQL, optional low cost support is available for a price that is $4 cheaper than MySQL Enterprise Gold. Migration from MySQL to DB2 should be pretty simple as DB2 and MySQL are very compatible. For those using dynamic language frameworks like Ruby on Rails and Zend Framework or Django it should be a no-brainer. You can expect very good performance (better than what you have with MySQL) and loads of new function (like hybrid engine with full XML support) that are simply not available in MySQL. The best part is that you are guaranteed not to be left in a lurch. DB2 Express-C is exactly the same code base as the other DB2 products. This means that you have 100% guarantee that any application written to work with DB2 Express-C will continue to work if you were to upgrade your database server to a higher value edition of DB2 for Linux, Unix and Windows. In many cases you would not even have to install another edition but should just be able to activate a new license.
You can wait and see if Oracle will put MySQL on death row or you can get ahead of the curve and start to plan for the alternatives like getting in touch with alternative MySQL support providers or take a look at DB2 Express-C and see if that is the right way for you to go.
Subscribe
To get an automatic feed of all future posts subscribe here, or to receive them via email go here. You can also follow me on Twitter.
MySQL will never die. Lets wait 3 years and we will see who is right.
We can use current version of mysql even without new features for years. It perfectly fits its primary deployement space – php web apps.
Hello Vlhka,
Everything would eventually depend on the Product Strategy that will help Oracle not the community @ large.
MySQL is only good @ powering websites? I think my favorite DB2 which I have using for almost 4 years now, is dfntly more capable of handling large data structures when compared to MySQL.
Cheers!
Anil Mahadev
IBM Data Champion
My point was that MySQL as an Oracle product is a non-starter. MySQL as an open source project probably has legs and companies like Percona and Monty Program will keep it alive.
At the same time, there is no harm in looking at what else is out there and DB2 Express-C is a great choice. It is FREE and it has a ton of technology that is absolutely top shelf. So, why not take a peek.
Your post overlooks one thing, Virtual Iron is almost virtually (excuse the pun) identical to Oracle VM. The Even the Oracle press release states this, and indicates features not currently available in Oracle’s own product will be rolled into a future release:
“Oracle has completed a thorough review of the Virtual Iron product portfolio, and intends to fully integrate Virtual Iron technology with Oracle VM, Oracle’s server virtualization and management product.”
This is very similar to what has been undertaken in the past with similar PeopleSoft and BEA products. However such a move cannot be viewed as an indicator of MySQL’s future. Things would be different if Oracle already had a commercial offering of MySQL, but there is a marked difference between it and their existing database products.
The point I wanted to make is that Oracle gave 11 days for existing customers to get new licenses and then it cut them loose. Product lines do get consolidated especially after acquisitions but 11 days at the end of the second quarter. This is just shafting VI customers which Oracle certainly does not care about. Who cares if Oracle VM is similar to Virtual Iron. If you work for an enterprise you have to rely on vendor stability. In other words, your company (procurement, CSO, CIO, CTO etc.) all make a commitment and you don’t just say “oh, Oracle press release says that they’ve concluded that Oracle VM is similar to what we use so we are OK”. You have a board of directors to answer to, you have auditors who may have certified your environment etc. When oracle pulls a stunt like that you loose your job and your resume does not look very hot either.
>Things would be different if Oracle already had a commercial offering of MySQL, but there is a marked difference between it and their existing database products.
I can assure you there is a lot less difference between Oracle database and MySQL than there was between Oracle VM and VI. The biggest selling point of MySQL Enterprise is that “Deploy an unlimited number of MySQL Enterprise Servers for the price of a single CPU of Oracle Enterprise Edition”. This comes right from http://www.mysql.com/products/enterprise/unlimited.html i.e. MySQL Enterprise Unlimited website. Do you really think this is still going to be around after the acquisition completes? Come on, this webpage is already next to impossible to find through normal site navigation of mysql.com.
Build your infrastructure over a closed-source software is madness. Luckily MySQL can and will be forked if Oracle tries to kill it.
MySQL is already forked. This is exactly my point. If you want to go with open source, than you are probably already resigned to the fact that there is no such thing as MySQL from Oracle and already looking for MySQL the way it used to be i.e. open source MySQL with support from small vendors like Percona etc. I don’t agree at all with the comment that “building IT infrastructure over closed source” is madness. MySQL may be open source but that is not its big attraction. Majority are drawn to it because it is free not because it is open source. last year the total number of contributions that MySQL accepted from the community was less than 40. Average Joe or Jane does not want to fiddle with the database kernel. They just want some free database for their web site.
As much as you are plugging DB2-express it is still crippleware compared to MySQL or PostgreSQL, even with the paid subscription. Not that you even provide the link or mention that.
It is in a similar product space as Oracle Express, Microsoft SQL server Express, or even Microsoft Access. They are all crippled products with upgrade path into their normal enterprise products.
That will be appropriate for some businesses but it is not the market which MySQL and MySQL Enterprise are in.
There is nothing stopping the MySQL forks from expanding into this market if Oracle kills it off, nor the companies providing the support from expanding to replace MySQL support.
Besides Oracle and IBM (via DB2/UDB) only want business from people who want to pay large amounts of dollars per core, their business costs don’t allow them to do anything else.
Paul, thanks for the comment. As far as plugging DB2 Express-C … guilty as charged. I do it every chance I get. I did update the post with a link to subscription. As I mentioned it in the post, it is $4 cheaper than MySQL Enterprise Gold i.e. $2995 per server per year.
DB2 Express-C is very different from both Oracle XE and Microsoft SQL Server Express. I agree that both of those are essentially crippled versions as they place a hard limit on the amount of data you can store. DB2 Express-C just like MySQL and PostgreSQL has no such limits. You can store as much data as your disk capacity will allow you. Want to run a 16TB database and have the storage for it? DB2 Express-C will do it. DB2 Express-C also has no limits on the size and type of the machine you can run it on. You can have a 32 or 64-bit machine with as much memory and as many CPUs or cores as you want. It may not use all of the resources you have on the machine but it will not restrict you in choosing your hardware or OS. And as I said in my post, any application you develop for DB2 Express-C is guaranteed to run on other editions of DB2 should you need more power.
As far as only wanting the business from people who want to pay lots of money per core … IBM is in business (as was MySQL AB) and we don’t apologize for it. With DB2 Express-C IBM has taken a bold approach and followed the same business model that MySQL uses i.e. help people get started by providing free licenses and help them with optional low cost support if and when they need it. I am sorry that you see this as a sinister plot to somehow extort money from unsuspecting customers. We have had DB2 Express-C available free of charge since February 2006. We have been very upfront and very clear on all terms and conditions. We have no small print. We provide this product for free use in development, production and even for redistribution. You can use it any which way you see fit. If you wan tot redistribute as part of your application, we just want to know but there are no royalties. It really is that simple.
The way DB2 manages users — via the OS — is a large hurdle for people coming from MySql (or any other DBMS for what that matters). DB2 express will never flourish before they provide user management like MySql and MS Sql have.
Ivo, I agree, not having user identity managed in the database is a bit of a let down for MySQL people. However, the direction in the database industry is to do exactly what DB2 does i.e. rely on an external security subsystem. This is driven by enterprise requirements to consolidate user identity management instead of having each and every application provide their own. I realize that this is an overkill for many startups and small businesses but our enterprise users are seriously paranoid (its a good thing) about security. To that end, DB2 has a pluggable authentication mechanism. By default it uses the operating system authentication but also has plug-ins for Kerberos, LDAP, Active Direcotry etc. This is actually supper convenient as it provides single sign-on capability. If you really want to, we do had people write plug-ins for authenticating using user credentials stored in hte database. So, if you want to use the same authentication model as MySQL you can.
ORACLE has not been successful breaking into the small/medium enterprises in the same way that MySQL and SQL Server has.
On one hand MySQL could be used as a route into those companies, on the other there is no natural upgrade path from MySQL to ORACLE.
Until ORACLE’s intentions are clear I wouldn’t want to be buying an Enterprise license beyond the next 12 months.
I am still not convinced that MySQL5.x is robust enough for enterprise level work, particularly in light of Monty Wideneus’s blog and even more so in the light of the Sun response to it. “We focus on features and timeliness over quality”. Whoopee the old M$ approach all over again!
Epsilon, this is exactly my point i.e. buying MySQL Enterprise at this point in time is a really dicey proposition. You are better off going to independents if you have support requirements.
Anil,
>MySQL is only good @ powering websites? I think my favorite DB2 which I have using for almost 4 years now, is dfntly more capable of handling large data structures when compared to MySQL.
People dont run databases, they run applications. How many PHP web applications supports DB2? Drupal, PHPnuke, PHPBB ? No. How many hosting providers supports DB2? None.
You are using DB2 on web, but you are statistic error in comparsion with people using LAMP stack.
>IBM Data Champion
Highschool Queen
@ Vlhlka
It might be true that MySQL trumps DB2 in web usage currently. IBM was definitely late to the game. But do not underestimate the power of DB2 here. They have been vetted in high-transaction rates in banking and insurance industries(both OLTP and DW).
I believe that as awareness grows of a DB2 free-edition coupled with high performance and a possible MySQL mishandling by Oracle, DB2 will eventually emerge winners (As long as IBM keeps innovation going along with a community outreach).
As an aside, I recently read an interesting article on DB2 vs MySQL performance. Its not all-encompassing, but still interesting:
(http://antoniocangiano.com/2009/06/05/do-androids-count-electric-sheep-with-db2-or-mysql/)
I’m not sure I’d even want my database handling identity management – I’d rather have the app, or AD, or LDAP handle it. This way you don’t have to maintain two sets of the same users (and sometimes the insane Sarbanes-Oxley rules that comes with that).
Usually setting up a connection user and an admin user for the app is good enough for most situations I’ve seen.
SQL Server, for example, has the option to do user management through database users, or an OS/AD user, and it makes life much easier to just let the Active Directory handle it.
I’m sure there must be a few pluggable PHP libraries out there already which provide some sort of application driven security mechanism for MySQL. You could couple that with a simple user table and you should be good to go.
I know the IBM crew is working hard to get DB2 free into the MySQL space. As mentioned, they were really late to the game. It’s hard to copy the success of an organic effort like MySQL/PHP from the corporate side, but it does have a track record, after IBM pushed Linux out from the hobby world and into the enterprise.
Ivo,
I would have to strongly disagree with you on user mgmt and mysql. This is a strong plus for db2 not a detractor. If I have 80 nodes for mysql, which I do, I have to create the user on each instance where as in db2 I can do it once in NIS or LDAP. DB2 is way ahead of the curve in that regard. Also I believe M$ does this too… so this isn’t really that novel of a concept. From a security perspective mysql is a nightmare to manage as users are by default allowed to have no pass. This makes the auditors very nervous indeed. Having done extensive work with both databases I can tell you this is one of MySQL weakest features.
kusu, Paul, Epsilon –
DB2Express-C will have some features like snapshots/online backups/easy index maintenance/an optimizer that works pretty well. These things are truly clumsy in MySQL (even in 5.x) on large tables (20-30GB+ tables and even on small tables in many cases). The fact that everybody thinks MySQL is cheap is simply confounding to me. You have recreate the wheel everyday with mgmt scripts and coding around the database. I would argue that this use of human resources is much more expensive than simply purchasing a reliable product.
@Vlhka,
It is ok if you do not feel a technology suits your needs.
There is no room for commenting on people’s titles + even giving them nicknames.
This is a public forum and trust that the moderator will take appropriate steps in not allowing abusive comments.
Peace!
Vlhka, I think it would be just stupid to argue that MySQL is not the king of LAMP stack. It is and no other database will ever displace it not PostgreSQL, not Oracle and not DB2. However, Some of the applications that you mention like Drupal, mediawiki etc. do run on DB2 and there are people who need/want to run these apps on DB2.
In my post, I was not trying to argue about MySQL’s ability to survive. I am sure it will lie as an open source project. My point was that MySQL Enterprise i.e. a supported version of MySQL that is meant for enterprise usage (not PHPNuke or PHPBB) has a much less certain future. The very fact that open source has already forked MySQL in several directions makes this future of the “Enterprise” even more uncertain. When you select technology for a large company, initial cost of license is rarely the number one concern. When you pick a technology, the investment you make in the ecosystem around that technology is many times that of the license cost. This is why vendor stability is so important to enterprise tech purchasers. One of the arguments when SUN was buying MySQL was that having a company like SUN behind MySQL would give enterprise customers more confidence when purchasing MySQL. I think that the Oracle acquisition of SUN has done the exact opposite. The overlap of MySQL with Oracle’s main business is making many people very nervous. And if the way Oracle killed VI, these customers have a good reason to be nervous.
DB2 Express-C ?? DB2 Express-C ?? No one uses that and why would they? Practically no one uses DB2 why would they be using DB2 Express-C . I guess it could be nice as all of the hosting companies provide access to DB2 Express-C … no wait none of them do, most of them haven’t even heard of it.
Runxc, I think when you say that “practically no one uses DB2” you are referring to building small web projects, porno sites, blogs etc. I agree completely, hardly anyone ever does these with DB2. Though you can now get Mediawiki, Drupal and a host of others on DB2 but finding an ISP that will host DB2 for you is a challenge. Thankfully the world is moving away from traditional ISP infrastructure and in to the Cloud Computing and DB2 is great and easy there.
Anyway, I digress. DB2 is the second largest DBMS by revenue behind only Oracle and well ahead of what think “everybody uses”. I’ll bet you that you use DB2 hundreds of times a day and may never even know it. When you pay for something with your credit card, your payment is cleared through DB2 and when you make a call on your cell, the telephone company processes much of the accounting for that call in DB2 as it probably does for fraud detection. Shipped a package with a courier lately? Yeap, you used DB2. Went on a Caribbean cruise, bought some fresh produce delivered by a trucking company, got health insurance claim paid? You get the point. Oh, and by the way, there are a few porn sites and on-line gambling outfits but we will not talk about those, they are a bit shy when it comes to publicity.
“DB2 is the second largest DBMS by revenue behind only Oracle… ”
I have to digress. AFAIK, IBM is the second company by DBMS revenues, next to Oracle. IBM does not publish exact numbers how much of the revenue is due to DB2 LUW, DB2 on mainframes, Informix, IMS… I’ve been curious, asking such info from some IBM guys, to no result. Informix is a jewel in IBM’s DBMS portfolio that is often neglected, even by IBM folks.
I don’t want to say anything but positive about DB2 Express-C as a product itself. It Deserves all the promotion it gets. If only there were similar Informix offering…
Darko, you are right, IBM does tot publish separate revenue numbers by product and anybody in position to know can not disclose it for variety of reasons including legal.
On the other hand, I would not be running afoul of the law or the IBM rules if I told you and everyone else who follows this blog that DB2 is indeed a major contributor to the overall IBM DBMS revenue. I am glad that you like IDS. It really is a terrific DBMS. The “neglect” statement is often expressed by some IDS customers however, I disagree with it. IBM has had a higher level of investment in to IDS than what Informix invested to it prior to acquisition. IDS has had more frequent and more feature reach releases since the acquisition so the product is far from neglected … it is actually doing very well.