Lots of Discussion on the backup plugin

The messages going back and forth on the “close sourcing” or paying to make use of some “plugins” /addons within MySQL are flying still! And in my opinion this is getting way out of hand. < DISCLAIMER >I work for Sun/MySQL as a senior consultant but claim no special knowledge or agenda here. The comments here are simply my opinions. </ DISCLAIMER> I can not help but look at all the hubbub and laugh a little bit. The vast majority of what I have read seems way off base (more of a he/she posted this, fifth hand retelling instead of from official channels). But there is a lot of FUD floating around, and I keep seeing more and more legit news sources picking up on this “massive change to licensing”, which In my opinion it really is a non-story… because there is no massive change.


What is the issue?

Basically it was announced that some backup plugins would only be made available to Enterprise customers. This follows the same distribution model of the enterprise monitoring tool, which had much less negative press when it was released as an enterprise only add-on. This is the way I see it ( not the views of sun and or their management:) . MySQL has many vendors who produce “add-ons”. Currently the majority of these add-ons come in the form of storage engines. Some of these storage engines are not freely available. None of the storage engine vendors get the type of grief that MySQL has gotten since they announced the new backup engine. Where was the outcry at the UC for open access to the kickfire ( I think I saw they had their own storage engine) or the infobright engine? From what I have seen the new “backup plugin/engine/api” is the same type of thing. Companies outside of MySQL will have access to the same plugin architecture. I am sure backup vendors may even start writing their own proprietary backup plugins. Is the outcry just because its MySQL releasing the backup plugins? If MySQL/Sun released a propitiatory storage engine would their be this level of outcry? If a new different company released Matt’s awesome backup plugin and did not open source it would their be any outcry? In fact if I made backup software (like Zmanda does) I may look into developing my own uber cool plugin that hooked in better to my product, which would hopefully help me sell more licenses. Many vendors create both open source and proprietary software ( and the plugins may not be proprietary anyways). This is nothing new. Why do we tolerate this in some cases but not others?


Does this really hurt the community?

Well some people can say the community is really hurt by this, but I do not buy it. Why? First off nothing is being taken away from anyone. In fact this is a new pluggable backup architecture, so additional features will be added that are available to all. This means the worst case is you still use the same backup methods you use today. Even better, there sounds like there may even be new and better backup features that will be freely available ( nothing official has been announced yet). Another even better case is the community of non-enterprise subscribers will get to have an API in order to build their own backup solution ( which maybe better then the pay one). So for those not skilled enough to build their own backup plugin they can simply wait for someone else to develop one and release it open-source or pay for the subscription. This is a big win for customers who need whatever functionality the plugins will provide. So the outcome is actually better then most people realize, with really no downside operationally ( worst case is the same, but with more options to develop something new, or finally pay for something with more functionality).


So who is really hurt from this?

On the surface other companies offering support and services for MySQL could be hurt. I know if I owned my own shop offering support and services I may look at this as a bad thing. Some people make a decent living helping people tune, troubleshoot, and design MySQL installations. I know if I was out on my own doing those services I would see this in the add-on features as a direct threat to my business( althought maybe not a huge one). Let em explain: If these plugins make purchasing enterprise more attractive, and more people start purchasing an enterprise license I may see a direct reduction in my business. The reduction is a side effect…if a company purchases enterprise to get the plugin, they will also get a support contract. Companies with an existing support contract will be less likely to purchase that support from an outside source. So this reduces one of their potential streams of revenue. On the flip side those companies may have more options for market penetration when additional backup plugins from third parties and community sources become more readily available. Hey backup plugin developer could be a hot job when 6.0 is released!


What about code quality?

One of the arguments I have seen over and over again is that code quality of the backup plugins is going to suffer because the code is not out in the open for everyone to see. This falls to Sun/MySQL to ensure the quality of the plugins is top notch. Look to the evolution of the enterprise monitor. In my experience the enterprise tool teams has done a great job. All the customers who I have worked with absolutely love it, and I have not heard of any quality issues ( completely unscientific benchmark I know ). The bottom line is these are backup plugins… if they suck, no one will install the plugin and believe me when I say everyone will hear about it on slashdot the day its released:) But here it is up to the community both paying and non-paying customers to keep things honest. As with all software, if there is a better alternative customers  will eventually gravitate toward it. So if a community released backup plugin beats the snot out of the “enterprise plugin” that’s great! This means new opportunities outside of Sun/MySQL for all.


Well that is my 2 cents! Since everyone who has a blog chimed in on this I thought I would.




This entry was posted in mysql, rant. Bookmark the permalink.