SUN, MySQL needs you to be bold, go for the cloud!

SUN wants to deal with MySQL like a small open source companies would: offering a few proprietary components on top of a free platform. This is so wrong: not only does it anger the community but it doesn't bring SUN anywhere near a solution that would leverage the long tail of MySQL adopters. How could you invest $1B in a company and subsequently foster such petty thoughts? Time to think differently.

MySQL is a special case in the open source world.  There are really 3 kinds of markets defined by those 200M+ downloads (correction: 6M-12M active installs depending on sources):

1. Zillions of small shops and start-ups
2. Medium-sized companies growing out from the above ones
3. Large companies whose technology guys are stealth-using MySQL in every corner

Good,  My SQL is all over the place but:

1. The long tail doesn’t have much money and those companies already have a few LAMP guys on board
2. The middle tier is going to be harvested efficiently by Microsoft and Oracle all-powerful sales force with which SUN cannot really compete
   
3. The F1,000s don’t even know they are using SQL and if they do they are more likely to be convinced to use ORCL IBM or MSFT light products depending on the shop who owns them
   

Bummer.  Time to ask the real question: what does the market need?   

1. Small guys have specialty needs and process issues:  scalability, security,  back-up and limited cash:  incremental  investments only.
2. The low-end middle tier needs reasons and incentive to stick to  MySQL instead of switching to MSFT/ORCL. They need help turning into "real", 1,000+ people companies.
3. The high-end middle tier needs help building their own IT infrastructure
4. The F1,000s need Enterprise integrated MySQL support with their Linux deployments.
   

Now, what is the right model for MySQL? As pointed out by Matt in a really great article, some lessons have to be pondered about. Here is a suggestion to SUN:

1.    Drop the non-scalable part of the business and pass it on to Redhat

Carve an agreement  with RedHat: allow them to use the trademark, pass on those download listings to RedHat salesforce, have Red Hat help distribute the service below.

2.    Focus on the long tail

• Offer what nobody else can offer: a convenient, on-demand  MySQL service coupling  virtualization, load-balancing, security and automatic back-up. Think Google apps or Amazon Rainbow (sounds nicer than SUN's cloud) but from a much larger install base, with a standard/universal interface (SQL) and therefore with NO fear of lock-in. People can prototype on MySQL and the day after they can launch; Who would say no to speed-to-market?
  
  

• Collect money on growing companies: (i.e. consuming too much bandwidth/Disk space/CPU) but keep it a much more interesting valued prop than that of Oracle or Microsoft (shouldn’t be that hard). That way, growing companies won’t be tempted to switch to much more expensive technology from the big guys: outsmarts Oracle and Microsoft on every single deal.

• Offer a transparent transition to enterprise data centers: Some companies want everything in-house: offer them an easy way to transparently transition   that you will  support with RedHat. Up sell your CISC servers at the same time. Use SUN virtualization technology to make the process completely seamless.

• Keep the community alive, separated and growing: give back everything developed in house and doing so further the strength of MySQL. With this model you can be bold, your force lies in your infrastructure. That way, many more developers will be inclined to contribute than with this narrow dual model, more suited for smaller companies and that often piss off entire communities.

As a side effect, you will have an opportunity to put to work all of those soon-to-be obsolete RISC architecture-based processors into a useful ...cloud. Remember  when I told you that MySQL needed a new storage engine? Well, here is an opportunity to give it a cloud-based storage engine.

Let's summarize:

• A F1,000 presence through RedHat
• A happy and thriving community
• A dual cloud/grid ultra-storage infrastructure à la Amazon/Google (and also a good way to keep them in check)
• An opportunity to leverage obsolete RISC-based processors
• A new MySQL strorage engine pluging directly into SUN cloud/infrastructure
• A clear way to outsmart MSFT and ORCL on their midle-range market
• An opportunity to sell (CISC) servers  and to generate scalable, recurring service revenues
• Cost? Maybe 10-15% of MySQL acquisition
• Time to Market? SUN has the ability to start a closed trial in just a few months.
• Wall street? Happy as a ClamAV without TrendMicro. It's going to be useful when those bad  numbers are going to be released during Q4

Developers -and yes, open source developers too- have a great respect for SUN.  Guide them instead of disappointing them. Go one step further: enroll the community to develop your cloud/grid: build a data center and give it to the community. Let's make it an ever lasting Summer of (distributed) code!

 

Why would  SUN largest investment in the future  be dealt with yet another model, another idea of the past?

 

Article quoted or cited by:

• Matt Asay on   c|net Open Road
• Mike Masnick on Techdirt
• Stephen O'Grady on Redmunk  

 

Comments

[Marten Mickos]

Pierre,

Wow - thanks for your suggestions! These are great ideas and something that I hope we will be working on. It shows how the power of open source (+ open debate) can extend to business models.

As for the business model, I realise we may have created a controversy of some sort in our ecosystem. Perhaps we are right or perhaps we are wrong in this particular topic. But we think it is vital that open source companies experiment with various business models. The industry is too young to settle on something that happens to look good at the moment. At MySQL, a key reason for our ability over the years to produce a ton of GPL software (and open some closed products) is that we have always had a functioning business model. As the marke environment changes, we will want to try out new models.

In some postings I have encouraged people to suggest a better business model to us if they are unhappy with what we are doing. Well, you now suggested one, and I very much appreciate it.

Perhaps we should set up an open session (or unconference) on this topic for people to brainstorm together?

Marten
formerly CEO of MySQL AB, now SVP at Sun

[Eduardo Pelegri-Llopart]

Hi Mårten,

An unconference/open session would be very interesting. I think products, market segments and adoption points lead to different best strategies for revenue generation (and sustainability of the community and development).

Definitively not a one-size-fits all.

- eduard/o

[Brian Aker]

Hi!

A few thoughts on the cloud:

A database is just a small piece of the infrastructure here. With a lot of work we could create a AWS/GApps environment, but the database is just one piece of a much larger infrastructure. A standalone database is pretty dull (and thin layers have been done by groups like DabbleDB). If you want backup just to S3... I am pretty sure Zmanda has that.

For cloud apps right now I see a dividing line, those who want an open stack, and those who want a hosted apps environment (which BTW Oracle has done for some time now). The Amazon model is really a mix. You can host LAMP stacks on EC2 entirely self contained, or you can make use of the other AWS services. The GApps approach is an entirely hosted application environment.

The question for anyone getting into this side of the business, aka both MySQL and others, is which is the track to follow?

Cheers,
-Brian

[Michael Widenius]

Some quick comments:

It was not Sun's idea to do closed extensions. This idea came from the MySQL management and Sun was caught by surprise by it.

I think it's quite clear that the reason Sun bought MySQL AB was to get an even stronger open source offering and also to learn from the things that MySQL AB has done right. Sun is into growing the MySQL business, not in milking the cow business.

Some of the things that you mention in the long tail is what Sun is already doing. Project carolina uses the programming languages Java and Perl, a computing grid and virtualization to provide a hosted platform for scalable Internet services. Currently this uses PostgreSQL but it will shortly also offer MySQL support.

I totally agree that it's vital for MySQL to keep the community alive! In my recent talk at the 2008 MySQL conference I pointed out that we need to change the development model to make it more community driven.

There are very few successful projects with closed source components that have an active community; The community usually doesn't want to participate in creating an software offering that they can't use them self!

You can find more things about this topic on my blog at: http://monty-says.blogspot.com

Monty
Former CTO of MySQL AB, now with Sun's CTO Organisation

[Mike Masnick]

Hey there,

Btw, you link to my article at Techdirt, but you call it TechCrunch. Please note that those are two totally separate and different sites.

MTG: oooops, 'course, fixed. :)

[verhuizen]

Great article and interesting site! You've got some worthwhile ideas. Thanks!

[Jeff Paul Big League Players Club]

Hi, Great post! I felt great reading your blog post. I’m working with my friend in a small internet marketing business as a web developer. When I’m free I go around for some IT info

[Cheap Health Insurance]

MySQL is the world's most popular open source database software, with more than 100 million copies of its software downloaded or distributed throughout its history and taking a MySQL certification training program will help us to learn all about the program.

[China Sources Procurement]

In recent years, the CMDBA certification has become a global standard for many successful IT companies. There has been an increase in the importance of achieving our CMDBA MySQL certification today.

[Industrial Product Design]

Hello,

Many of the world's largest and fastest-growing organizations use MySQL to save time and money powering their high volume Web sites, critical business systems and packaged software and knowing how to properly use it will help our future.

Thanks,
Industrial Product Design(http://www.streamline-alliance.com/)

[Chinese Market Research]

Hi,

PhpMySql provided by WingSix, actually allows a database to share space so we can use it for more than one program. Although it is not ideal, we should be able to have our forum and our blog in separate tables of the same database without asking our provider to give us more databases.

Thanks,
Chinese Market Research(http://www.streamline-markets.com/)