Are load balancers disruptive to NetApp software licensing fees?
One of our customers has been asking questions about whether they should take the plunge into NetApp metro clusters. And at first glance the technology looks pretty interesting. But then we remembered that Jeff Browning, the former NetApp engineering big shot, mentioned that NetApp clusters run as a Java Virtual machine, and he implied that it is not a very elegant solution. It works though. As Jeff says...
NetApp's version: In the event of failure, the OS environment representing the failed member of the cluster reboots inside a virtual machine environment on the surviving member, running under, you guessed it, Java. Then this virtual machine Java thingy takes ownership of its storage objects, and continues whatever it was doing. With lots and lots of overhead. Goody!
When we reviewed the costs of the licenses that are required and compared that to the costs of reconfiguring the network to use Load balancers and two sites with snapmirror, the costs seem to favor a load balanced scenario.
NetApp likes to talk about disruptive technologies, but in some ways an old technology like load balancing might be disruptive to their pricing model.
Can NetApp improve its cluster technology and lower its costs? Can an old technology like Load balancing and some network engineering produce higher reliability at a lower cost? Disruptive technology works in strange ways, especially when customers are driven to reduce the cost of their storage infrastructures.