I've had the same need - a previous posted mentioned AutoSys, which for all of its ugly faults gave my last employer a very robust job scheduling platform that I found very reliable.
I've been looking (waiting?) for an open source equivalent. What we really need is something like Condor and Globus, ala the NSF Cluster Toolkit, with a cron interface (cluster-centric solutions have great features like redirection of STDOUT and STDERR, but don't have the ability to schedule a job for later execution.) Java Workflow systems have the kind of business logic you would want, but lack the cross-platform job execution and STDOUT redirection needed. Good luck finding something.
Re:it is VERY trollish
on
The Faded Sun
·
· Score: 1
A lot of the discussion here seems to focus on the merits of Linux and such.
I don't think people would disagree that Linux has become a viable solution for low to mid-sized servers (4 CPUs or less), and can achieve scalability when implemented in a horizontal manner (e.g. as in web farms.)
But for high-volume, mission-critical databases you're in a different ballgame. You have several implementation options, namely one big database server in an HA configuration, or using Oracle 9i RAC to achieve a horizontally-scaled database tier across multiple machines. I think we would agree that the larger architecture benefits from Sun hardware. IMHO you'd be crazy not to use a Sun, HP, or IBM Unix variant due to their stability and maturity, and proprietary-but-reliable hardware.
However, scaling the database tier with 9i clusters running Linux makes sense - its cheap and designed to "work around" running on less reliable hardware. I am assuming the storage is SAN-based and is reliable in its own right.
Sun is in trouble because Intel has made the microprocessor a commodity. It costs at least $1 billion to develop a microprocessor. Intel can spend $2 billion, sell more volume to multiple vendors, and hence have a huge R&D and price advantage. Sun will have a hard time competing on that front, and it is the SPARC chip that I view as a key reason Suns are more expensive.
SPARC vs Intel performance aside, application architectures are changing to a model where CPUs and CPU power is a commodity. Sun's answer is N1, a very loosely defined vision that does not fit the bulk of the IT shops out there, but clearly articulates a vision where its own hardware is a commodity. Commodity these days means Intel.
Answer - migrate to Intel CPUs like everybody else, and differentiate yourself with super servers with all of the reliability and redundancy that makes Suns great servers in the first place. Using Intel CPUs does not mean dropping the same I/O architecture that make Suns great database servers. That does not mean dropping their great on-site support.
It just makes them Dell & Compaq/HP, with better gear and better service with hardware whose cost structure is in line with what the industry expects these days. I love Sun gear; I just can't justify spending three times as much in certain areas of the data center.
Hopefully they will make a change before the software industry does, as Linux/x86 has the momentum that could overtake them if they don't act fast.
This would be a great story - security engineer tracks down malicious hacker (Mitnick.) You got engineering, crime, and suspense. Embelis a bit and you're golden.
Slashdot Mirror
I've had the same need - a previous posted mentioned AutoSys, which for all of its ugly faults gave my last employer a very robust job scheduling platform that I found very reliable.
I've been looking (waiting?) for an open source equivalent. What we really need is something like Condor and Globus, ala the NSF Cluster Toolkit, with a cron interface (cluster-centric solutions have great features like redirection of STDOUT and STDERR, but don't have the ability to schedule a job for later execution.) Java Workflow systems have the kind of business logic you would want, but lack the cross-platform job execution and STDOUT redirection needed. Good luck finding something.
A lot of the discussion here seems to focus on the merits of Linux and such. I don't think people would disagree that Linux has become a viable solution for low to mid-sized servers (4 CPUs or less), and can achieve scalability when implemented in a horizontal manner (e.g. as in web farms.) But for high-volume, mission-critical databases you're in a different ballgame. You have several implementation options, namely one big database server in an HA configuration, or using Oracle 9i RAC to achieve a horizontally-scaled database tier across multiple machines. I think we would agree that the larger architecture benefits from Sun hardware. IMHO you'd be crazy not to use a Sun, HP, or IBM Unix variant due to their stability and maturity, and proprietary-but-reliable hardware. However, scaling the database tier with 9i clusters running Linux makes sense - its cheap and designed to "work around" running on less reliable hardware. I am assuming the storage is SAN-based and is reliable in its own right. Sun is in trouble because Intel has made the microprocessor a commodity. It costs at least $1 billion to develop a microprocessor. Intel can spend $2 billion, sell more volume to multiple vendors, and hence have a huge R&D and price advantage. Sun will have a hard time competing on that front, and it is the SPARC chip that I view as a key reason Suns are more expensive. SPARC vs Intel performance aside, application architectures are changing to a model where CPUs and CPU power is a commodity. Sun's answer is N1, a very loosely defined vision that does not fit the bulk of the IT shops out there, but clearly articulates a vision where its own hardware is a commodity. Commodity these days means Intel. Answer - migrate to Intel CPUs like everybody else, and differentiate yourself with super servers with all of the reliability and redundancy that makes Suns great servers in the first place. Using Intel CPUs does not mean dropping the same I/O architecture that make Suns great database servers. That does not mean dropping their great on-site support. It just makes them Dell & Compaq/HP, with better gear and better service with hardware whose cost structure is in line with what the industry expects these days. I love Sun gear; I just can't justify spending three times as much in certain areas of the data center. Hopefully they will make a change before the software industry does, as Linux/x86 has the momentum that could overtake them if they don't act fast.
This would be a great story - security engineer tracks down malicious hacker (Mitnick.) You got engineering, crime, and suspense. Embelis a bit and you're golden.