Brian Wong, one of Sun’s Distinguished Engineers spoke this morning and stated categorically that the “Storage [Market] is right to be disrupted”. He argued that the general purpose OS (such as Solaris) offers massive developer economies of scale, by which we mean operating system developer economics. He quoted an example of one of our disk controller operating systems which we have 31 developers and no community, where as for Solaris, Sun employs 1200 people with an extended community of tens of thousands. Even if fixing bugs was the only work that developers need to do, 31 is not a lot of people, but storage devices and hence their OS need to evolve to remain useful.
He claimed that there are a number of myths about the nature of a storage device operating system, the most prevalent of which is that it needs to be real time. Despite the fact that Solaris has a real time scheduler, Brian argued that storage doesn’t need real time.
Furthermore, Solaris is well positioned because as large drives and larger drives come onto the market, Sun’s portfolio of storage operating systems functionality which now includes ZFS, Solaris Cluster, the fault management architecture & SMF, together with the highly functional SAM-FS/QFS and Sun availability suit delivers storage functionality to storage administrators and architects. It may also act as very attractive platform for new entrants to the market as opensolaris is available under the Community Development and Distribution License which means that they do not inherit a duty to publish their innovations.
What with the industry leading science in tape devices, we have some interesting times ahead.
Originally posted on my sun/oracle blog, and re-published here, https://blog.davelevy.info in June 2016.