Tag Archives: Solaris

Performance misconceptions on storage networks

The piece of spinning Fe3O4 (ie rust) is by far the slowest piece of equipment in the IO stack. Heck, they didn’t invent SSD and Flash for nothing, right. To overcome the terrible latency, involved when a host system requests a block of data, there are numerous layers of software and hardware that try to reduce the impact of physical disk related drag.

One of the most important is using cache. Whether that is CPU L2/L3 cache, DRAM cache or some hardware buffering device in the host system or even huge caches in the storage subsystems. All these can, can will, be used by numerous layers of the IO stack as each cache-hit means it prevents fetching data from a disk. (As in intro into this post you might read one I’ve written over here which explains what happens where when a IO request reaches a disk.)

Continue reading

Oracle (and when the SUN doesn’t shine anymore"

I’m not hiding the fact I’ve been a SUN MicroSystems fan all my life. They had great products, great engineering philosophy and best of all great people who knew how to pick a potato. The problem was that they went down the same path as DEC (huhhh, who…???) DEC, Digital Equipment Corporation. One of those other fabulous engineering companies who fell as pray to the PC world due to their lack of marketing knowledge and sales strategies. Google around for that.

Oracle was by far the worst company to acquire SUN. They have a massively different company mindset which is 100% focussed on getting another boat for Uncle Larry “I want you.. o, no.. I want your money” and this went on to be a head-on collision with the SUN philosophy. Given the fact Oracle had a massive war-chest and SUN was struggling to keep afloat allowed them to get the entire SUN IP for a nickle and dime.

The worst thing for Oracle was that all of a sudden they inherited a hardware division with, be honest, great products but also a huge drag on sales numbers. (which was likely to be the reason for SUN’s struggling.) No easy way out here since product support and near term roadmap line-ups had to be fulfilled. Oracle is, has always been, and always will be, a software company so over the last couple of years you can already see that the majority of all hardware products are starved to death. Don’t expect any new developments here.

SUN was bought for two reasons: Java and Solaris. Well, only certain parts of Solaris. COMSTAR was one of them an ZFS the other. Java of course was the biggest fish since that piece of the pie runs in almost every device on the planet from cell-phones to toasters. ZFS allowed Oracle to create Exadata  and tailor this to very specific workloads. (Duhhhh…. lemme guess> Oracle Databases). The funny thing is that they almost give away this Exadata box since they know it only performs well with their database and this is were you start paying the big bucks.

So lets get back to what is left of SUN. SUN was also a very big supporter of the open source world. Projects like OpenOffice, Netbeans, GlasFish etc are all neglected by Oracle for them to die a certain death. OpenOffice (originally acquired by SUN as StarOffice had a really nice spin since some developers had absolutely no trust in Oracle anymore and “forked” the entire code branch in LibreOffice which is now the most actively maintained Office Suite outside M$ Office. Oracle is sidely hanging on to MySQL and allow some people to put some effort in this project. The reason is obvious. Its a stepping stone to one of Oracle own big bucks databases and suites so the biggest sponsoring is done on migration software from MySQL to OracleDB itself. If Oracle decides to pull the plug on MySQL it will be simply forked as well and continue under another name were Oracle has absolutely no insight and loses any business advantage. Don’t ever expect any Oracle IP going into MySQL. Larry needs a bigger boat.

Another product SUN “donated” to the open source world was OpenSolaris. A free (as in free beer) spin of SUN’s mainstream operating system. SUN’s intentions for OpenSolaris was to provide a free platform for developers to have easy access to Solaris. This would allow for more applications to become available and as such a larger ecosystem to live on into companies using those. The steppingstone to a revenue generating operating system for those applications would then be real easy. A similar fashion Microsoft has followed for quite a while. (Provide a real cheap consumer product for developers to hook onto and sell at a premium to companies). Unfortunately it wasn’t mend to be so as soon Oracle took over the OpenSolaris project was starved to death.

So when taking into account all things that happened with the SUN acquisition it is very sad to see that such great products and philosophy is butchered by pure greed. Many distinguished engineers  like James Gosling, Time Bray and Bryan Cantrill left immediately and many more followed. The entire Drizzle team resigned, as well as all of the jRuby engineers. In fact the only SUN blooded executive to stay was John Fowler who kept onto his hardware group.

In retrospect the only thing Oracle bought for that 5.6 billion dollars is Java which is a very heavy price for a piece of software (soon to become obsolete) and an empty shell.

This once more shows that great products will always lose against marketing, an effective salesforce and a money hungry CEO.

Don’t get me wrong, I’m not against someone making a fair chuck of money but effectively killing of an entire company and leaving so many people in the cold doesn’t really show any form of ethics. A good friend of Larry, the late Steve Jobs, had similar characteristics however he also had a heart for great products.

Regards,
Erwin

PS. The comment of Java becoming obsolete is because many major new web technologies are now being put in place to bridge the gap to Java. This includes semantics, document and extensive data control, device control etc. Within 5 years Java will likely have a serious competitor which allows developers to gain more freedom and interoperability than Java now can provide.