For those of you who are not aware, Oracle (database and enterprise software) bought Sun (JAVA and big iron servers) after the IBM deal fell apart. The San Jose Mercury News wrote that “Analysts said the deal could transform (Oracle) into a data center giant — and a new rival with IBM, Hewlett-Packard, and even Cisco.”
In my opinion, Oracle will be working on very tight integration between the two companies and will begin to sell a juiced up Oracle Database Engine against Oracle running on any other platform.
But HP has a weapon that is so secret that many sales reps are only dimly aware that it exists. I am referring to the HP NonStop server – the most scalable, available system on the planet. First designed in 1974 by a handful of ex-HP employees, the NonStop server is completely modularized and redundant at both the hardware and software level.
Scaling from 2 to 4080 Itanium processors with up to 127 TB (yes TB) of RAM, and driving petabytes of disk, this amazing beast runs many of the world’s most critical applications without drawing attention to itself.
The NonStop server supports an ANSI standard SQL database that runs on the metal. Thousands of parallel threads can insert, update, and query the database at the same time without impacting each other. When was the last time you ran a database query on a live loaded system?
NonStop servers are so stable that many customers don’t even know how to fire one up because they have been running since the customer engineer (CE) installed it. After the San Francisco earthquake a customer called in to ask how to fix their down system – and they meant that it fell down but was still processing transactions. You will never hear the question, “Have you rebooted your system” from a NonStop support agent.
The NonStop server is a workhorse with average downtime measured in minutes per year – that means seven nines. And if it does fail for any reason, it automatically restarts, ensures that the databases are consistent and continues from where it left off. For all intents, the NonStop server is a private cloud running within your data center.
As the system load increases, new processes will be instantiated and mapped into the application and when the load decreases they will be torn down to make room for other loads – without operator intervention. The system is self healing and even phones home to tell the CE what parts will be needed for repair. Customers which have chosen to ignore the extensive logging have been surprised to learn of a component failure when the engineer calls to set up a repair appointment.
Without special programming nor adding overhead to the application, the HP NonStop server can survive a catastrophic site failure without losing a single transaction in progress. And did I mention that consistent, recoverable backups are done while the application is running?
I seriously would be considering how to use the HP NonStop server to protect HP from the Oracle-Sun onslaught against Oracle on HP-UX. And don’t think that it won’t be coming, because Larry doesn’t like to lose.