Not joking, but betting that your business parallelizes wonderfully, so you can break up your transaction processing over a Beowulf cluster.
Nope. Only ever done one of those. It ran Fluent (on Redhat) and was for the Aerodynamics group for one of the Foruma One racing teams. The bulk of my experience comes initially from working for one of the big vendors, and then self employed as a consultant in banking, telcos, pharmaceutical and health. Some of the data crunching software typically deployed in these areas are: Ab Initio, Nucleus, Octopus and Oracle RAC.
Alas, large transaction processing tends to require something like a POWER or SPARC, to get 128 cores with a common locking architecture working on driving a large database.
Traditionally, yes. But times are changing as the top Intel Xeon and AMD chips are finding their way into systems that can easily deliver the horsepower to meet those needs. And memory capacity is no longer an issue either. Besides TPC performance is more sensitive to fast IO and lots of spindles for more IOPS than the difference between the fastest CPUs. Just take a look at the configs the system vendors put together to hit the top TPC benchmark results. The disks number in the thousands! And Oracle's fastest published results are on clusters of Xeon chipped systems.
Yeah, but the thing is that 32-CPU systems are incredibly niche. I've been involved in projects that delivered a number of systems of that size over the years and I can count on one hand the times they've been used as single 32-CPU systems. In virtually all cases they were hard partitioned down to 4, 8 and sometimes 16 cpu systems. And x86 is walking all over that market now. Next year when the Nehalem-EX chips ship, you'll get your 32 cores on a standard 4 socket server with twice as many threads. It just shoves the high end systems more and more into a tight corner. RAIDed memory is great, but that alone is not worth the premium that proprietary solutions are burdened with.
Same as with all proprietary high end solutions: customer ignorance. The customer goes to the vendors and says: "Here's my shopping list of business requirements. Please bid a solution that meets those needs". The vendor salesman (after wiping the drool from his/her chin) comes back with an Enterprise Class solution using propietary high end kit at the highest price the saleman thinks he/she can get away with to win the bid but beat off the competition. The whole thing is wrapped up in smoke and mirrors to make the customer feel valued and special with the assurance that they're getting the best in class. The whole thing is topped off with generous dollop of FUD dissing any other vendor solution. Things like: "The x86 space is too aggressive and its 3 year turnover cycle is bad for your business. Use our systems which have a 5 year life cycle and get a better return on your investment". Or here's another one: "Our chips are built with advanced RAS features. They're [self healing] and crash less often than x86". Oh and lets not forget that to buy one of their enterprise solutions, you usually also have to buy their proprietary enterprise OS and pay their enterprise software license fees at their inflated enterprise prices.
Not your fault, you're just not familiar with OS X version naming, which I'll admit is very strange compared to the rest of the industry and made no sense to me at all when I first saw it. They insist that Mac OS X should be pronounced Mac OS Ten, but it doesn't sound right when you say Mac OS Ten 10.6, so most people slip in a silent "version" and say Mac OS Ten, version 10.6. I guess after several years I just don't think about it any more.
In any case, the "10" part of the version number appears to be fixed in concrete. I don't think they're ever going to change it. Consequently 10.4, 10.5 and 10.6 are all major upgrades. Think of them as 10.4.0, 10.5.0 and 10.6.0 with many sub versions in between.
Why they don't version OS X the same way everyone else does I've no idea, but there you are.
Expecting a vendor to support double major version upgrades in one leap is taking a big risk on your part: most OS vendors don't support this. OS X is currently at 10.5.7, so even if you're on the latest 10.4 build, a leap to 10.6 is skipping forward 9 OS X releases in one go. Something is almost guaranteed to break. My guess is that the upgrade patch from 10.4 will be to go to 10.5.x first and then jump to 10.6. Either that or reinstall.
# Each guest needs its own kernel, so you need to allocate memory and disk space for all these kernels that are in fact identical
Actually this depends on your virtualization solution
No it doesn't. The parent is clearly talking/complaining about VMware, Xen, Kvm type virtualization, and guest OS instances for all those require their own kernel. He isn't talking about jails/container solutions (FreeBSD Jails, OpenVZ, Solaris Containers, etc) or none of his points would make any sense.
# A guest's filesystem is on a virtual block device, so it's hard to get at it without running some kind of fileserver on the guest
You can often mount the virtual disks in a HOST OS. No different to needing software to access multiple partitions. As long as the software is available, it's not as big an issue.
Not without shutting the guest down first. If you mount a filesystem on a disk/partition twice and that filesystem is not a specially designed cluster filesystem, and the two OS instances are not part of the same cluster, then you WILL get data corruption. The parent's point is valid !
You should have stopped at your list of what virtualization is good and not good for. You let yourself down after that.
Both of these would work great in a corporate level with a WAVE server running on the LAN, but when it goes global, those servers will be smokin'
So there's something new coming along that might require a beefy server to drive it. I can see the sales guys from {HP,IBM,Sun/Oracle} wiping the dribble off their chins already. Would you like a SAN with that Sir?
It will be useful to know (when its released) what size of server and how much bandwidth Google recommend to support # numbers of users.
For some people it already is. I know a couple of people (not me yet) who've decided that for home stuff they just don't need a full blown PC with Office software anymore and are using Google for everything now. Google Docs are perfectly adequate for writing letters and doing simple tasks that require a spreadsheet. They will only get more sophisticated, and internet bandwidth will keep on increasing, so this will only get better.
It's not centralized, they're making it Open Source and publishing the protocols and APIs. The idea is that every organisation can have their own wave server, in the same way that they have their own SMTP Email servers day. Also, they're anticipating that other organisations will write their own fully compatible, standards compliant Wave servers in competition to Google's.
Want to have a conversation with several people from different organisations (with wave servers) about a project, then open up a Wave for the project and just invite them using their email addresses. The intro vid was long, but they did cover this.
Thanks for the link, I'd not seen that. Just spent the last 1hr20 watching the vid. WOW, this answers my needs on so many levels. I look forward to the day when this is as ubiquitous as SMTP, and email can finally be put to rest.
Ah thanks. Wasn't sure how strong my memory was with the model numbers. It was a long time ago. After posting I had a good think and now believe it was the difference between an 11/780 and 11/785.
I hear you. I joined in 86 and stayed for 13 years. There were 130,000 employees at the start but it had that small company feel and was like working with family. Many of the friends I made then have lasted even till today. You just don't get that in large corporations now. Senior management are too focused on the stock market and shareholders to overly concern themselves with "social contracts" with their employees. Maybe places like Google are (or were) the exception - I don't know. I imagine Redhat would be a great place to work. Their employees seem very energised, enthusiastic and excited about the work they do. If I were ever tempted to drop out of the consultancy life and back into working for someone else, I think Redhat would be top of my wish list.
Hm, I didn't think we had it in the south, as I've never met or even heard of anyone else getting it apart from me. But a google search has turned up a few hits. There seems to have been an increase since 2005 due to rising temperatures, so it's pretty recent.
If memory serves me right, the only difference between a VAX-11/750 and a VAX-11/780 was an upgrade to the firmware floppy for one with all the CPU NO OPS taken out. And from what an engineer at the time told me, DEC stole the idea from IBM !
So vendors have been playing tricks like this since the late 80's.
This brings back memories. I went to the Spit Brook site once for 3 weeks in the mid 90's to spend 3 weeks with the DEC MLS+ engineering team. At the time that was the B1/CMW (Compartmented Mode Workstation) high security variant of Digital Unix. There were no training courses, so I had to go learn it from the horses mouth, so to speak, so I could support it when I got back home. I remember Spit Brook well for 3 reasons:
- Great atmosphere at the place. People were excited and enthusiastic about what they were doing. And I'd never seen such a collection of raw talent in one place before. Really bowled me over.
- It was in the middle of the biggest pine forest I'd seen in my life. Walking out the hotel in the morning I would just stop or 5 minutes and breath it all in. Never experienced air like that before, or since.
- I got invited to a cook out (had never heard that expression before) and while there I got attacked by this mahoosive black fly. I thought I'd managed to avoid getting bitten, but when I got back to the UK I discovered several strange looking bites. A red spot surrounded by a large white circle and a red ring around that. Only time in my life I've ever seen a UK doctor routing through a text book to work out what I had. He eventually diagnosed it as Lyme Disease. Apparently the fly picks it up from feeding on deer. We don't get it in the UK. A course of antibiotics shifted it.
Oh, and there was a 4th reason: Diane Lebel. I should never have left, or turned around and gone straight back. Enough said;-)
This reminds my of the futuristic OCtattoos (Organic Circuitry tattoos) that Peter Hamilton makes use of in his Commonwealth Saga stories.
They're described in the Wiki as:
OCTattoos (Organic Circuitry Tattoos) are also a major technological device. These are tattooed on the skin and resemble colourful, often metallic tattoos, and serve hundreds of purposes from transferring credits to serving as sensors. Their main function is to act as processors for other implants (which may function at reduced capacity if an OCTattoo is damaged).
Would be really neat to have this in our lifetime.
.. for stealing commercial software. No sympathy at all. If people want software for free with relative security and safety, then the proper way to go about it is to find an Open Source application that fits the need. Ya know, that's what Open Source is there for, duh !
I'm with you in that I hardly ever see a single spam email these days. There was a time when the Junk folder in my Mail app held a pretty consistent ~3000 spam emails. Today it just has 2 !
It's tempting to wonder why the spammers even bother anymore, except we know that they only do it because enough people respond to generate plenty of profit for them all.
So the only conclusion we can draw from this is that not all Mail services are created equal. Google is the king in my book, but there must be others that are lousy.
Who are the bad boys out there? They need to be named and shamed.
I don't work for HP, but my background comes from Digital, Compaq and working very closely with HP and its resellers for several years. I know exactly what you're talking about, and have seen it over and over.
If an Enterprise customer gives HP or one of its pet resellers a blank technical sheet and a list of business requirements, what the customer gets presented is a gold plated solution based on Itanium and HP-UX, with FC SAN storage underneath. It doesn't matter that the customer could get a faster solution at a fraction of the price using Proliants and Linux. They get what the salesman earns the most commission on, and what his overlords have tasked him to sell. And HP can't be blamed for that; they're in the business of making money. So naturally they're going to go with what generates the most profit.
So you're absolutely right that most Enterprise Linux solutions are driven from the customer side. Those customers that are in the know and have good people on their own team to advise them are able to direct HP (and by the sounds of it IBM too) to deliver a solution based on the technologies they want to use. Those who haven't got a clue will get spoon fed the most expensive dishes.
The biggest issue you need to overcome with FOSS projects in a business setting is supportability. For example, I'm on a project at the moment where I'm transitioning the customer from a proprietary unix solution onto multiple Oracle RAC clusters on Redhat; Oracle Application servers on Redhat; and Linux Virtual Server load balancing clusters, also on Redhat. This is fine, because the software stack from top to bottom is mainstream, supported by commercial vendors, and after I'm gone there is a well defined set of skills they can recruit against and train existing staff to replace me. Since getting here though I've discovered a few bespoke applications (developed in-house by people who have since left) written using Ruby on Rails. While the apps work well today, documentation is poor to non-existent, and no one is left now with skills to understand them, develop them if requirements change or support them. They aren't backed by a vendor, so if something goes wrong they're screwed. It's kind of their own fault: they gave free rain to someone who either wanted to do this stuff using his own favourite tools, or wanted a tick on his resume, instead of sticking with technologies in line with their core competencies. If you want to do something with Drupal for example, then make sure you're able to wrap it up in a support structure (from a vendor) that can give them the security they need. Another example: I convinced my current customer that switching to Zabbix for their server, application and network monitoring and alert needs would be a good thing, and they went for it. Why? Because while Zabbix is Open Source, it's also backed by a vendor (Zabbix) and they can buy a commercial support contract. In addition, being a FOSS project they could install and test it at no cost for as long as they like before making a decision and parting with their cash. So if you can tick all the boxes, you stand a much better chance of getting your ideas accepted.
And don't listen to anyone who tells you to sneak this stuff in through the back door. If it's under the radar then your employer is in for a nasty surprise if it goes wrong. And if it's business critical you'll find yourself pink slipped faster than you can blink.
Slashdot Mirror
Not joking, but betting that your business parallelizes wonderfully, so you can break up your transaction processing over a Beowulf cluster.
Nope. Only ever done one of those. It ran Fluent (on Redhat) and was for the Aerodynamics group for one of the Foruma One racing teams. The bulk of my experience comes initially from working for one of the big vendors, and then self employed as a consultant in banking, telcos, pharmaceutical and health. Some of the data crunching software typically deployed in these areas are: Ab Initio, Nucleus, Octopus and Oracle RAC.
Alas, large transaction processing tends to require something like a POWER or SPARC, to get 128 cores with a common locking architecture working on driving a large database.
Traditionally, yes. But times are changing as the top Intel Xeon and AMD chips are finding their way into systems that can easily deliver the horsepower to meet those needs. And memory capacity is no longer an issue either. Besides TPC performance is more sensitive to fast IO and lots of spindles for more IOPS than the difference between the fastest CPUs. Just take a look at the configs the system vendors put together to hit the top TPC benchmark results. The disks number in the thousands! And Oracle's fastest published results are on clusters of Xeon chipped systems.
Yeah, but the thing is that 32-CPU systems are incredibly niche. I've been involved in projects that delivered a number of systems of that size over the years and I can count on one hand the times they've been used as single 32-CPU systems. In virtually all cases they were hard partitioned down to 4, 8 and sometimes 16 cpu systems. And x86 is walking all over that market now. Next year when the Nehalem-EX chips ship, you'll get your 32 cores on a standard 4 socket server with twice as many threads. It just shoves the high end systems more and more into a tight corner. RAIDed memory is great, but that alone is not worth the premium that proprietary solutions are burdened with.
Same as with all proprietary high end solutions: customer ignorance. The customer goes to the vendors and says: "Here's my shopping list of business requirements. Please bid a solution that meets those needs". The vendor salesman (after wiping the drool from his/her chin) comes back with an Enterprise Class solution using propietary high end kit at the highest price the saleman thinks he/she can get away with to win the bid but beat off the competition. The whole thing is wrapped up in smoke and mirrors to make the customer feel valued and special with the assurance that they're getting the best in class. The whole thing is topped off with generous dollop of FUD dissing any other vendor solution. Things like: "The x86 space is too aggressive and its 3 year turnover cycle is bad for your business. Use our systems which have a 5 year life cycle and get a better return on your investment". Or here's another one: "Our chips are built with advanced RAS features. They're [self healing] and crash less often than x86". Oh and lets not forget that to buy one of their enterprise solutions, you usually also have to buy their proprietary enterprise OS and pay their enterprise software license fees at their inflated enterprise prices.
Perhaps you think I'm joking !
Have a sniff around on ebay. I'm sure you can pick up a copy of 10.5 for a lot less than that now, and prices will drop even more when 10.6 comes out.
Not your fault, you're just not familiar with OS X version naming, which I'll admit is very strange compared to the rest of the industry and made no sense to me at all when I first saw it. They insist that Mac OS X should be pronounced Mac OS Ten, but it doesn't sound right when you say Mac OS Ten 10.6, so most people slip in a silent "version" and say Mac OS Ten, version 10.6. I guess after several years I just don't think about it any more.
In any case, the "10" part of the version number appears to be fixed in concrete. I don't think they're ever going to change it. Consequently 10.4, 10.5 and 10.6 are all major upgrades. Think of them as 10.4.0, 10.5.0 and 10.6.0 with many sub versions in between.
Why they don't version OS X the same way everyone else does I've no idea, but there you are.
Dear anonymous crap-hard. He's not crying, he's laughing at you: we all are.
The Tomtom demo didn't fail, and that was the most impressive third-party peripheral of the bunch,
Expecting a vendor to support double major version upgrades in one leap is taking a big risk on your part: most OS vendors don't support this. OS X is currently at 10.5.7, so even if you're on the latest 10.4 build, a leap to 10.6 is skipping forward 9 OS X releases in one go. Something is almost guaranteed to break. My guess is that the upgrade patch from 10.4 will be to go to 10.5.x first and then jump to 10.6. Either that or reinstall.
No it doesn't. The parent is clearly talking/complaining about VMware, Xen, Kvm type virtualization, and guest OS instances for all those require their own kernel. He isn't talking about jails/container solutions (FreeBSD Jails, OpenVZ, Solaris Containers, etc) or none of his points would make any sense.
Not without shutting the guest down first. If you mount a filesystem on a disk/partition twice and that filesystem is not a specially designed cluster filesystem, and the two OS instances are not part of the same cluster, then you WILL get data corruption. The parent's point is valid !
You should have stopped at your list of what virtualization is good and not good for. You let yourself down after that.
So there's something new coming along that might require a beefy server to drive it. I can see the sales guys from {HP,IBM,Sun/Oracle} wiping the dribble off their chins already. Would you like a SAN with that Sir?
It will be useful to know (when its released) what size of server and how much bandwidth Google recommend to support # numbers of users.
For some people it already is. I know a couple of people (not me yet) who've decided that for home stuff they just don't need a full blown PC with Office software anymore and are using Google for everything now. Google Docs are perfectly adequate for writing letters and doing simple tasks that require a spreadsheet. They will only get more sophisticated, and internet bandwidth will keep on increasing, so this will only get better.
It's not centralized, they're making it Open Source and publishing the protocols and APIs. The idea is that every organisation can have their own wave server, in the same way that they have their own SMTP Email servers day. Also, they're anticipating that other organisations will write their own fully compatible, standards compliant Wave servers in competition to Google's.
Want to have a conversation with several people from different organisations (with wave servers) about a project, then open up a Wave for the project and just invite them using their email addresses. The intro vid was long, but they did cover this.
Thanks for the link, I'd not seen that. Just spent the last 1hr20 watching the vid. WOW, this answers my needs on so many levels. I look forward to the day when this is as ubiquitous as SMTP, and email can finally be put to rest.
Ah thanks. Wasn't sure how strong my memory was with the model numbers. It was a long time ago. After posting I had a good think and now believe it was the difference between an 11/780 and 11/785.
I hear you. I joined in 86 and stayed for 13 years. There were 130,000 employees at the start but it had that small company feel and was like working with family. Many of the friends I made then have lasted even till today. You just don't get that in large corporations now. Senior management are too focused on the stock market and shareholders to overly concern themselves with "social contracts" with their employees. Maybe places like Google are (or were) the exception - I don't know. I imagine Redhat would be a great place to work. Their employees seem very energised, enthusiastic and excited about the work they do. If I were ever tempted to drop out of the consultancy life and back into working for someone else, I think Redhat would be top of my wish list.
Hm, I didn't think we had it in the south, as I've never met or even heard of anyone else getting it apart from me. But a google search has turned up a few hits. There seems to have been an increase since 2005 due to rising temperatures, so it's pretty recent.
If memory serves me right, the only difference between a VAX-11/750 and a VAX-11/780 was an upgrade to the firmware floppy for one with all the CPU NO OPS taken out. And from what an engineer at the time told me, DEC stole the idea from IBM !
So vendors have been playing tricks like this since the late 80's.
Thanks. I'll try this when I get back home at the weekend. 9.04 is unusable for me at the moment.
This brings back memories. I went to the Spit Brook site once for 3 weeks in the mid 90's to spend 3 weeks with the DEC MLS+ engineering team. At the time that was the B1/CMW (Compartmented Mode Workstation) high security variant of Digital Unix. There were no training courses, so I had to go learn it from the horses mouth, so to speak, so I could support it when I got back home. I remember Spit Brook well for 3 reasons:
- Great atmosphere at the place. People were excited and enthusiastic about what they were doing. And I'd never seen such a collection of raw talent in one place before. Really bowled me over.
- It was in the middle of the biggest pine forest I'd seen in my life. Walking out the hotel in the morning I would just stop or 5 minutes and breath it all in. Never experienced air like that before, or since.
- I got invited to a cook out (had never heard that expression before) and while there I got attacked by this mahoosive black fly. I thought I'd managed to avoid getting bitten, but when I got back to the UK I discovered several strange looking bites. A red spot surrounded by a large white circle and a red ring around that. Only time in my life I've ever seen a UK doctor routing through a text book to work out what I had. He eventually diagnosed it as Lyme Disease. Apparently the fly picks it up from feeding on deer. We don't get it in the UK. A course of antibiotics shifted it.
Oh, and there was a 4th reason: Diane Lebel. I should never have left, or turned around and gone straight back. Enough said ;-)
This reminds my of the futuristic OCtattoos (Organic Circuitry tattoos) that Peter Hamilton makes use of in his Commonwealth Saga stories.
They're described in the Wiki as:
Would be really neat to have this in our lifetime.
I'm with you in that I hardly ever see a single spam email these days. There was a time when the Junk folder in my Mail app held a pretty consistent ~3000 spam emails. Today it just has 2 !
It's tempting to wonder why the spammers even bother anymore, except we know that they only do it because enough people respond to generate plenty of profit for them all.
So the only conclusion we can draw from this is that not all Mail services are created equal. Google is the king in my book, but there must be others that are lousy.
Who are the bad boys out there? They need to be named and shamed.
I don't work for HP, but my background comes from Digital, Compaq and working very closely with HP and its resellers for several years. I know exactly what you're talking about, and have seen it over and over.
If an Enterprise customer gives HP or one of its pet resellers a blank technical sheet and a list of business requirements, what the customer gets presented is a gold plated solution based on Itanium and HP-UX, with FC SAN storage underneath. It doesn't matter that the customer could get a faster solution at a fraction of the price using Proliants and Linux. They get what the salesman earns the most commission on, and what his overlords have tasked him to sell. And HP can't be blamed for that; they're in the business of making money. So naturally they're going to go with what generates the most profit.
So you're absolutely right that most Enterprise Linux solutions are driven from the customer side. Those customers that are in the know and have good people on their own team to advise them are able to direct HP (and by the sounds of it IBM too) to deliver a solution based on the technologies they want to use. Those who haven't got a clue will get spoon fed the most expensive dishes.
I know, I know. Long day and I'm tired.
The biggest issue you need to overcome with FOSS projects in a business setting is supportability. For example, I'm on a project at the moment where I'm transitioning the customer from a proprietary unix solution onto multiple Oracle RAC clusters on Redhat; Oracle Application servers on Redhat; and Linux Virtual Server load balancing clusters, also on Redhat. This is fine, because the software stack from top to bottom is mainstream, supported by commercial vendors, and after I'm gone there is a well defined set of skills they can recruit against and train existing staff to replace me. Since getting here though I've discovered a few bespoke applications (developed in-house by people who have since left) written using Ruby on Rails. While the apps work well today, documentation is poor to non-existent, and no one is left now with skills to understand them, develop them if requirements change or support them. They aren't backed by a vendor, so if something goes wrong they're screwed. It's kind of their own fault: they gave free rain to someone who either wanted to do this stuff using his own favourite tools, or wanted a tick on his resume, instead of sticking with technologies in line with their core competencies. If you want to do something with Drupal for example, then make sure you're able to wrap it up in a support structure (from a vendor) that can give them the security they need. Another example: I convinced my current customer that switching to Zabbix for their server, application and network monitoring and alert needs would be a good thing, and they went for it. Why? Because while Zabbix is Open Source, it's also backed by a vendor (Zabbix) and they can buy a commercial support contract. In addition, being a FOSS project they could install and test it at no cost for as long as they like before making a decision and parting with their cash. So if you can tick all the boxes, you stand a much better chance of getting your ideas accepted.
And don't listen to anyone who tells you to sneak this stuff in through the back door. If it's under the radar then your employer is in for a nasty surprise if it goes wrong. And if it's business critical you'll find yourself pink slipped faster than you can blink.