Fujitsu Siemens plans Linux innovation
1 April 2005 | 0
Success for Linux needs to involve original applications — the open source operating system shouldn’t resort to merely copying what others have done, according to one hardware manufacturer supporting its adoption.
‘The key to Linux’s success is not only something that repeats the possibilities in other areas such as Windows, but that it is a platform for innovation in its own right,’ said Joseph Reger, chief technology officer with Fujitsu Siemens Computers.
Strong applications available on Linux will also enhance its appeal beyond the fact that it is available at a relatively low cost, Reger added. He pointed to the Flexframe for MySAP concept, a recent development that FSC undertook with SAP and Network Appliance. Such a setup normally requires a number of servers, with a particular MySAP service allocated to a physical server. Now, these services are virtualised.
‘Where Linux plays a role is that for all these servers in the MySAP suite, they run an identical copy of the operating system. Keeping the operating system in the NetApp filer and is handed over to the blades on an on-demand basis. Everything boots from the very same image. It is very stable and very available,’ he said. This allows customers to rip out and replace blades as needed, and have their load reallocated to available components. Fujitsu Siemens previewed this capability at last month’s CeBit trade show in Germany.
Reger added that this sort of ‘net boot’ capability is currently only available on Linux. ‘This is very clearly something of a differentiation at the moment,’ he said. ‘It’s important that there are things happening on Linux that aren’t happening elsewhere. Linux shouldn’t be looking at what others are doing and just copying it.’
‘Three to four years from now, Linux will live in heterogeneous environments. Find ways of integrating Linux into mixed environments: that’s the best approach. We don’t view it as diluting anything. I think it’s key that the Linux market doesn’t splinter, like Unix did in the 80s. That was not good for anyone.’
In tandem with that, Fujitsu Siemens’ recently launched Linux strategy sees the computer maker offering the operating systems across its portfolio of hardware platforms, from mainframe and Sparc-based systems to Intel servers. ‘It’s very clear to us, we don’t expect the commercial ISVs and middleware vendors to support those platforms the same as they do for Linux on Intel,’ Reger told ComputerScope. ‘The ISV support for applications and middleware will be very strong on Intel platforms and less so on others.’
He claimed that FSC’s strategy was ‘very different’ from that of rival IBM, which he said was based on the belief that high end Linux will happen on big systems only. ‘We don’t believe that high end Linux will happen on RISC or mainframe architecture. In our estimation, there will be a high end Linux market on high end Intel systems competing with traditional Linux systems and the two will sit side by side in data centres.
In the meantime, there is still a commercial case for Linux on Unix hardware, Reger pointed out. ‘If you look at Solaris today, there is a high degree of compatibility with Linux. Linux applications on Intel can be recompiled and run on Solaris.’
In spite of all the current excitement about Linux, it will not be the single operating system in an environment, Reger suggested. ‘Integration with Unix and Windows is key.’