Open Sauce Software
Tasty titbits from people using Linux and other open source software in business.
Thursday 20 December 2007, 4:57 PM
An Open Source Census?
After musing earlier htis week on the difficulty of knowing how much open source software there is, here comes a possible answer.
The Open Source Census, launched by enterprise open source company
The Open Source Census, launched by enterprise open source company
Tuesday 18 December 2007, 10:31 AM
Open source -- apps lag OS?
Forrester research has found that open source applications are doing less well than the Linux operating system in enterprises.
On one level that's obvious - There's a lot more awareness of Linux than there is of applications (unless you count firefox, of course). And for many people Linux is the start.
"For Linux and all other open source software, service/support and security remain the two top concerns among enterprises. If open source is going to continue to be an important tactical tool, the open source ecosystem must address these concerns," Forrester said, in the study, which spoke to 2500 enterprise and SMB "decision makers" in Europe and the US.
Why the distinction between "Linux" and "open source"? What seems to have happened is this. Forrester decided to treat Linux separately this time, so it's not a new result, just something that was revealed by new treatment of the figures.
More interestingly, the survey says "open source is not a high priority among strategic software initiatives, appearing to be more of a tactic..."
In other words, users aren't adopting open source for idealistic reasons, but because of what it does. And what it does is increasingly strategic. They use products and don't care if they are open source.
And they may not even know they are open source. Some users might think they are steering away from open source because they perceive a risk there - but are they?
"Forrester has reason to believe that many application development professionals are not aware of how much open source software is contained within commercially licensed products they are currently using.
"If you are using Java, then chances are that somewhere in your platform there is open source. Being aware of its presence will help you leverage the larger collection of open source software more effectively."
On one level that's obvious - There's a lot more awareness of Linux than there is of applications (unless you count firefox, of course). And for many people Linux is the start.
"For Linux and all other open source software, service/support and security remain the two top concerns among enterprises. If open source is going to continue to be an important tactical tool, the open source ecosystem must address these concerns," Forrester said, in the study, which spoke to 2500 enterprise and SMB "decision makers" in Europe and the US.
Why the distinction between "Linux" and "open source"? What seems to have happened is this. Forrester decided to treat Linux separately this time, so it's not a new result, just something that was revealed by new treatment of the figures.
More interestingly, the survey says "open source is not a high priority among strategic software initiatives, appearing to be more of a tactic..."
In other words, users aren't adopting open source for idealistic reasons, but because of what it does. And what it does is increasingly strategic. They use products and don't care if they are open source.
And they may not even know they are open source. Some users might think they are steering away from open source because they perceive a risk there - but are they?
"Forrester has reason to believe that many application development professionals are not aware of how much open source software is contained within commercially licensed products they are currently using.
"If you are using Java, then chances are that somewhere in your platform there is open source. Being aware of its presence will help you leverage the larger collection of open source software more effectively."
Thursday 13 December 2007, 4:58 PM
Open source routers? Don't hold your breath
Hey, I just heard that Cisco's routers are going open source! That's exciting, isn't it! No it's not, because it's not true.
The words "open" and "router" created a small flurry this week, as Juniper "opened up" its routers, and Cisco seemed to make a counter-promise to open its routers too. But don't be fooled. Router software is definitely one of the most obvious targets for open source - but it is hard to see many signs of progress.
Routers really could benefit from open source. At the moment, users are paying their network vendors to re-invent the wheel, duplicating the same router software functions on different families of hardware, just because those router vendors want to keep control, and keep their margins high.
Software costs could be slashed, new features could be added more easily, and new hardware brought on-stream in an open source world. In web servers, open source can lead the world, why not in routers?
Routers are tightly bundled with their operating software, but they are all programmed with more or less the same style of command line interface, copied from Cisco's IOS. The vendors' OS code itself is rigorously protected - no-one knows how much money Cisco got from Huawei, for instance, when the Chinese network vendor admitted copying IOS code.
This week's announcements had nothing to do with open source. Juniper is allowing third parties to develop applications for its JUNOS operating system, but only those third parties that it approves of. That's a long way from open source. It's approaching the level of open-ness we've always had with DOS and Windows.
And Cisco's promise is less than that - it's simply vapour, designed to make the company look good while its analyst event is on.
Open source software could cut the cost of routers in half, says Vyatta Networks, the best-known open source router company. It could bring benefits to Cisco and Juniper in the form of better applications,
Vyatta uses the eXtensible Open source Routing Platform XORP, which has been plugging away for five years now, on general purpose Intel hardware. And why not? We think of routers as needing special purpose hardware, but remember, they all trace their lineage back to general purpose hardware, as early models ran on 1980s workhorses, like the DEC PDP 11.
Vyatta is encouraging its users to post "Dear John" letters to Cisco's John Chambers, explaining why they dumped his kit. It's also set up a two-tier distribution similar to Red Hat, but unless we're missing it, we haven't seen the explosion of open source routing that Vyatta hoped for.
It's possible, of course, that these competing "open" announcements from Juniper and Cisco might be the start of something. They could be a sign that the big guys are starting to feel the competition but, as we saw five years ago with Microsoft's shared source, there's a mighty long way to go between a company acknowledging the potential of opening up, and actually doing anything real about it.
The words "open" and "router" created a small flurry this week, as Juniper "opened up" its routers, and Cisco seemed to make a counter-promise to open its routers too. But don't be fooled. Router software is definitely one of the most obvious targets for open source - but it is hard to see many signs of progress.
Routers really could benefit from open source. At the moment, users are paying their network vendors to re-invent the wheel, duplicating the same router software functions on different families of hardware, just because those router vendors want to keep control, and keep their margins high.
Software costs could be slashed, new features could be added more easily, and new hardware brought on-stream in an open source world. In web servers, open source can lead the world, why not in routers?
Routers are tightly bundled with their operating software, but they are all programmed with more or less the same style of command line interface, copied from Cisco's IOS. The vendors' OS code itself is rigorously protected - no-one knows how much money Cisco got from Huawei, for instance, when the Chinese network vendor admitted copying IOS code.
This week's announcements had nothing to do with open source. Juniper is allowing third parties to develop applications for its JUNOS operating system, but only those third parties that it approves of. That's a long way from open source. It's approaching the level of open-ness we've always had with DOS and Windows.
And Cisco's promise is less than that - it's simply vapour, designed to make the company look good while its analyst event is on.
Open source software could cut the cost of routers in half, says Vyatta Networks, the best-known open source router company. It could bring benefits to Cisco and Juniper in the form of better applications,
Vyatta uses the eXtensible Open source Routing Platform XORP, which has been plugging away for five years now, on general purpose Intel hardware. And why not? We think of routers as needing special purpose hardware, but remember, they all trace their lineage back to general purpose hardware, as early models ran on 1980s workhorses, like the DEC PDP 11.
Vyatta is encouraging its users to post "Dear John" letters to Cisco's John Chambers, explaining why they dumped his kit. It's also set up a two-tier distribution similar to Red Hat, but unless we're missing it, we haven't seen the explosion of open source routing that Vyatta hoped for.
It's possible, of course, that these competing "open" announcements from Juniper and Cisco might be the start of something. They could be a sign that the big guys are starting to feel the competition but, as we saw five years ago with Microsoft's shared source, there's a mighty long way to go between a company acknowledging the potential of opening up, and actually doing anything real about it.
Friday 7 December 2007, 1:53 PM
Novell hits at Red Hat's real-time rant
Red Hat complained this week that Novell hadn't contributed enough to real time Linux, and was launching code developed by Red Hat that should still be in beta - Novell has responded.
"Novell is shipping tested and enterprise-hardened Linux with real time capabilities," says Kevin Barney over at Novell's Open PR blog. "Just because Red Hat is again late to market (see enterprise Linux desktop, Xen virtualization, etc.) doesn’t mean [SUSE] Linux contains “beta code.” SUSE Linux Enterprise Real Time is already run in production environments, providing real services to real customers like Thomson Financial."
Furthermore, Novell has contributed "a vast amount of code" to real time Linux, says Barney, along with others, and points to the real time community mailing list for proof.
Two issues arise from all this.
1. According to Barney, some sites were so confused about the nature of open source, they misreported the whole thing - saying that Red Hat was objecting to Novell's use of the code, something Novell is perfectly entitled to do under the open source licence. We don't have to explain what open source is all over again, do we?
2. What do we mean by real time? A couple of people have suggested to me that what real time kernels, like Lynx have hitherto been specialised beasts - are Red Hat and Novell actually offering something that's just a (major) step along the spectrum towards predictable response time?
"Novell is shipping tested and enterprise-hardened Linux with real time capabilities," says Kevin Barney over at Novell's Open PR blog. "Just because Red Hat is again late to market (see enterprise Linux desktop, Xen virtualization, etc.) doesn’t mean [SUSE] Linux contains “beta code.” SUSE Linux Enterprise Real Time is already run in production environments, providing real services to real customers like Thomson Financial."
Furthermore, Novell has contributed "a vast amount of code" to real time Linux, says Barney, along with others, and points to the real time community mailing list for proof.
Two issues arise from all this.
1. According to Barney, some sites were so confused about the nature of open source, they misreported the whole thing - saying that Red Hat was objecting to Novell's use of the code, something Novell is perfectly entitled to do under the open source licence. We don't have to explain what open source is all over again, do we?
2. What do we mean by real time? A couple of people have suggested to me that what real time kernels, like Lynx have hitherto been specialised beasts - are Red Hat and Novell actually offering something that's just a (major) step along the spectrum towards predictable response time?
Tuesday 4 December 2007, 5:28 PM
Vista - so bad, they pay you to take it?
We've noted the lack of fanfare around Vista's unhappy first birthday, but here's a sign of just how bad things are - online stores pay you to take it.
Look at Lenovo laptops on Dabs.com (but don't buy there - see below). An R61i with Vista Home Premium costs £400. R61is with XP Pro cost £567 or £575.
Over at eBuyer, an R61 costs with Vista Home Premium costs £401.61, while R61is with XP Pro start at £586.87 (prices include VAT).
I'm not sure if the detailed specifications are exactly comparable, but it looks like the Vista machines are about £170 to £180 cheaper. In other words, online stores are effectively paying you money to take Vista.
Obviously, the reason is that this is Vista Home Premium, the cut- down version of the new OS, compared with XP Professional, the business grade version of XP. But it's an eye-opener to realise that any version of the new operating system is considered to be £170 worse than the standard version of XP.
I found myself speaking to someone at Dabs' customer service (a difficult achievement in itself, part of the process of cancelling an order they failed to deliver - and very much, as they say, another story).
"Why not buy a Vista laptop and an XP professional disk?" she asked. Possibly because the £86 XP disk is an OEM version without long-term support, and a full consumer version of XP seems to cost £235.
But wait a minute. That price is almost the same as the £239 Dabs charges for Vista Business.
So, the actual value of all Microsoft's work to deliver Vista? A fiver.
Unless we should be comparing XP Pro with the pinnacle of Vista Ultimate, retailing at £324, or £178 for the upgrade. If we are, then that really is an admission of failure....
Look at Lenovo laptops on Dabs.com (but don't buy there - see below). An R61i with Vista Home Premium costs £400. R61is with XP Pro cost £567 or £575.
Over at eBuyer, an R61 costs with Vista Home Premium costs £401.61, while R61is with XP Pro start at £586.87 (prices include VAT).
I'm not sure if the detailed specifications are exactly comparable, but it looks like the Vista machines are about £170 to £180 cheaper. In other words, online stores are effectively paying you money to take Vista.
Obviously, the reason is that this is Vista Home Premium, the cut- down version of the new OS, compared with XP Professional, the business grade version of XP. But it's an eye-opener to realise that any version of the new operating system is considered to be £170 worse than the standard version of XP.
I found myself speaking to someone at Dabs' customer service (a difficult achievement in itself, part of the process of cancelling an order they failed to deliver - and very much, as they say, another story).
"Why not buy a Vista laptop and an XP professional disk?" she asked. Possibly because the £86 XP disk is an OEM version without long-term support, and a full consumer version of XP seems to cost £235.
But wait a minute. That price is almost the same as the £239 Dabs charges for Vista Business.
So, the actual value of all Microsoft's work to deliver Vista? A fiver.
Unless we should be comparing XP Pro with the pinnacle of Vista Ultimate, retailing at £324, or £178 for the upgrade. If we are, then that really is an admission of failure....
Previous
