The next Linux kernel will feature improved input/output and virtualization capabilities, says Tim Witham, CTO
of Open Source Development Labs. Prior to taking the stage at the LinuxWorld Conference & Expo San Francisco, Witham offered a preview of new kernel features and summed up the wins scored by Linux 2.6. The current state of the kernel is the topic of his conference session Tuesday.
What technologies will be coming to the kernel soon, and how can IT shops use them?
Tim Witham: Improved I/O for RDBMS (relational database management system) workloads, enhanced multi-path I/O for availability, and, of course, the virtualization giving VM to the COTS market segment.
What were Linux 2.6's primary technological wins?
Witham: Enabling virtualization is a big win as it allows IT shops to start their development cycles for a technology they will be looking at deploying within the next year or so. There has been lots of good work done with regard to system scalability, memory management, disk I/O, process and thread scalability. Also, work done for availability, like a greatly improved multi-path I/O [were victories].
Have any of the vulnerabilities found been major setbacks?
Witham: I can't recall any big items. They have tended to be small and quickly fixed.
Is kernel development focused on the needs of corporate/business/enterprise IT?
Witham: Many of the kernel developers are employed by corporate/business/enterprise IT companies. This is particularly true for the major distributions, where they have direct insight in the kernel development process. Because of this, you can bet that most kernel developers have a focus that will allow their companies to succeed in their individual markets.
How does Linux on the server stack up against Windows on the server today?
Witham: Linux is currently the fastest growing operating system (OS) in terms of market segment share. I regularly talk to IT shops that are adding hundreds of Linux servers at a time while keeping all other OSes flat, except for some specialized applications. In short, they have made the switch; Linux is the default for deploying applications. Other OSes are limited to the special case situations.
Can you compare the kernel/distribution approach to that of Microsoft with Windows or a Unix variant?
Witham: In many ways it is no different than the developer releases versus the packaged distributions for a closed source operating system. Sure, you can track all of the internal development by asking for lots of NDA updates, but that isn't what you should be looking to for your planning.
If you are a commercial shop, you should be talking to the company that you are going to get your SLA (service level agreement) from and asking for their roadmap and doing your planning based upon that. The difference is that if you want, you can then look at the open development process to check the progress without anybody from a marketing department filtering it for your consumption.
Many CIOs tell us that they don't have enough in-house expertise to use Linux and open source software extensively. Could you offer some training tips that won't break their budgets?
Witham: The first thing to do would be to ask their employees if they have Linux expertise. If they do, I think they will find a core of Linux expertise in their own company. Also, many community colleges and universities have courses in Linux system administration.
In your talks with business leaders, what concerns about Linux and open source come up most often? How do you respond to those concerns?
Witham: They ask me the same sorts of questions that I used to get when I worked for a proprietary systems vendor: "How do I make this piece better," and "How do you feel about X or Y?" Almost every situation seems to have a different set of main issues, so I don't really see one general area that comes up more than others.