This looks pretty good for the network performance alone, plus the RAID controller is pretty nifty. If it works nicely under Linux (and nVidia's site appears to have drivers) it'll make a good chipset for a low end server. I like what I've seen of the new AMD stuff, particularly their new system design (Hypertransport etc) . Had a presentation from Sun recently on their coming Opteron lineup and that looks pretty sweet too (looks like IBM is offering them as well).
Agreed. I tend to run Sun equipment because it's rock bloody solid and just works. The other reason is Solaris. I find it to be among the more reliable and easy to admin server O/S's out there. I just wish DEC was still around. Loved those Alpha systems (yes, yes I know HP sorta still has 'em, but Compaq effectively killed the platform years ago). If that makes me behind the times or whatever, I couldn't care less -- I'll take my uptime, scalability and reliability thanks. Would love to have a G5 on my desk though:)
Patching Solaris really isn't that hard. You either download the patch you need and install it or simply get the whole recommended cluster and run the built in script. I find it no more difficult than any other platform. If there are pre-requisite patches, then the web page for the patch on Sunsolve tells you what they are. Pretty simple if you ask me. Granted, OSX is pretty nice too and patching my iBook is fairly trivial.
As to your other point, I have rarely seen Solaris patches outright break stuff to the point of rendering applications unusable. I normally feel pretty confident that they'll work as advertised.
I would argue that studio costs have come down significantly in the last 10 years, since DAW's (Digital Audio Workstations) came on the market. It is now possible to build up a semi pro recording rig for only a few thousand dollars now, whereas before the analogue multitrack tape machines were often tens of thousands of dollars by themselves. Leaving aside the analogue vs digital debate (pointless IMHO), I think it's possible to produce very good recordings fairly cheaply with a PC/MAC based solution these days.
The thing to remember is that the recording software is only a drop in the bucket compared to the cost of the other equipment. Good mics and outboard equipment is where the majority of the money goes (not to mention preamps and mixing desks if your needs extend that far).
Of course, a good pair ears and some skill is the most valuable asset in a studio:)
Sun has had the ability to do multiple system images on the same box for a while, but they've always been hardware partitioning only. The 4800/6800/12k/15k allowed you to run different domains on the same system, so long as you had the right combo of CPU and I/O boards. This was great if you had one of those systems, but not so hot it you had a workgroup level system (e.g. E450 or V880). I'm glad to see they've put software partitioning in the O/S so I can take a mid range system and chop it up into separate pieces. AIX and HP-UX have been able to do the software side thing for a while (but not the dedicated hardware piece, I believe).
This will help with consolidation and utilisation on existing machines, I think.
I just finished reading this book and agree with the recommendation. It tells the story of what happened at that remarkable facility over a period of about 10 years. It was certainly rather illuminating. A must, I think, for anybody interested in the history of computing.
The odd thing for me was how advanced systems like the Alto were and that so many of the innovations it contained within took so long to come to market.
Didn't say anything about Sun in particular. There aren't many outfits that do massive scalability like SGI, and those new Altix servers are very nice (we have a few ourselves). They aren't the only solution on the market and neither are SPARC's. BTW, solaris scales very niceley on big iron, just look at it on the top end FujitsuSeimens systems.
I think you rather proved my point that shared memory systems still have great performance characteristics.
Ahh, it depends!! Some apps require the sub processes to be able to communicate with each other, preferably over a very high speed interconnect. This is where NUMA and SMP are handy. They allow near linear scaling of performance as you add processors. This is not always the case with clusters. I can take most multithreaded apps and get predictable scalability on an SMP system, but not always on a cluster.The other thing to consider is if you need a single memory image, versus fragmenting over many nodes. Do you need cache coherency to go with it?
To give an example, the company I work for writes a piece of software that does reservoir simulation. It is parallel capable and can be run on cluster or SMP systems. An IBM p690 will run the simulation around double the speed of a PC cluster (using Myrinet and Xeons at nearly double the clock speed) with the same number of total processors. This tells me that for this particular application, I would strongly consider using SMP iron.
When it comes to databases, you have to look at how the engine is clustered. Is it "shared all" (where the nodes in a cluster access the entire database with shared storage) or "shared nothing" (where the database is broken down into chunks with each node having its own piece). Last time I checked, Oracle was shared all and DB/2 and SQLSERVER were shared nothing. This makes a huge difference in how you set them up.
The other factor when it comes to SMP vs Clustering, is maintenace and management of the nodes. Even very well run clusters require more admin time than single systems.
It's actually a good sign when upper management wants to talk. You really can take the opportunity to work through issues that affect the department and the business. The best IT groups form a partnership with their user base and use it to work on issues together.
My advice:
1, Confront the real issues. On both sides. There is no point in trying to dance around major problems if they are affecting the organisation.
2, Be honest. The old cliche about it being the best policy is actually pretty true. It is the best policy. Face the brutal truth and it sets you free in a funny kinda way. Once you do this, you can set about fixing the problem.
3, Be tactful. Ranting, raving and whining get you nowhere, particularly with management. I have found that airing concerns, but couched in a way that show the impact on the organisation as a whole works well. Upper level managers can relate to it that way.
4, Actually show concern about the direction of the business. This goes along with the whole partnership deal. It engenders a bit of trust, which can be a huge help to you later when you ask them to take a chance on a big risky project you want money/time/people for.
5, Listen to what they have to say. This is obvious, I know, but it's amazing how much you learn when actively listen to where their pain is. They'll love you forever if you can show exactly how to ease it.
6, Have a game plan. Think of what you want to talk about first, even if it's in broad terms. You don't waste people's time that way.
I would treat this as a golden opportunity to make progress on issues and generally gain a better understanding of their business and them of yours. You may never get a better chance, particularly in an informal situation like dinner. If at the end of the day nothing changes, you at least gave it a shot.
Nothing has really changed in terms of building a successful business, IMHO. Providing a product/service that your customers want to buy at a price they are prepared to pay and keeping them happy will increase your chances of success. Of course, there's more to it than that, but in effect that's it.
A book I found very useful is "Good to Great" by Jim Collins, which covers why some companies fall apart and some succeed. It came down to good leadership, top quality people, understanding of core business, discipline and facing the hard realities. Worth a read for anyone involved in private enterprise.
All capitalism does is allow people to act in their own rational self interest, which they would do anyway as IMHO that is a fundamental human trait. At under a capitalist system, they can be compensated for their own work and then have the freedom to spend the fruits of their labours in any way they choose (in the pure capitalist system, anyway). That doesn't sound like exploitation of people to me.
Now don't get me wrong, there are corporations and individuals who have done terrible things to both the environment and other people in the name of maximising profits. Those involved should go to jail, not be the catalyst for some structural change in the economy. Most folks are decent and law abiding, so why should their private property rights be abolished because of the actions of a few individuals?
In any case, I would argue that Communism caused a great deal more damage to people's rights and the environment than capitalism. How many people were killed in the USSR and China in the last century? You can add religious fundamentalism to the list as well.
What you are describing is not quite libertarianism. For it to qualify, there would have to be some system to enforce contracts and preserve private property rights. Without a legislature and a court system, that cannot happen. What we see in places like Somalia is coercion and force used to resolve private disputes, which is the antithesis of a libertarian (IMO) point of view.
Pure any "-ism" can never work anyway, whether it's Libertarianism, Communism et al, as you're never going to get a country full of people to agree on the best way to run it. Better, IMHO, to have the representative system as in most countries, where people get to have a say periodically and elect those whom they think will most effectively cary out their wishes.
Agreed. Good on Sun for stepping up and providing an alternative.
I think the reason why Excel has become such a standard is that it's a bloody good spreadsheet in it's own right. I couldn't say the same for Word and Powerpoint (which have gotten harder to use and more bloated as time has gone on), but Excel beats other mainstream spreasheets I have used hands down. There are obviously more specialised products out there that do a great job in certain disciplines (such as SAS or Mathematica) but Excel is great for general purpose stuff like I do...
In my experience, those who earn those kinds of salaries are rarely in an easily replacable position. Furthermore, they are often given whichever tools give them the ability to do their jobs most efficiently. If that deviates slightly from a "standard" then so be it. The 150k a year salary might look high on the P&L, but if they are good they add substantially more value than that to the organisation (i.e in new sales, new ideas or increased efficiency). Their pay starts to look cheap then....
A very good point. I work in IT, but I actually really like what I do. Furthermore, I can't imagine not doing it (well, maybe unless I was a musician). I think it's important to have job you like, as you're at work anywhere from 8 - 12 hours a day. Added to that, if you enjoy it, you're generally better at it, get paid more because of your skills and probably be happier in other parts of your life ('cause you're not bitching about work all the time).
I am constantly amazed by folks who get into a career simply because it pays well, but are otherwise miserable. These same people often times will bitch about their work but not do anything about it. I saw this a lot when I worked in the Government sector. I guess they become so entrenched that they see no way out. Distressingly, I also see this in education from time to time, particularly high school teachers, but some times at tertiary level.
...then I would consider building a SAN with replication. High end storage solutions using HDS and/or EMC gear fix this problem by enabling remote block for block copy of data between identical arrays. Veritas also makes a product called Volume Replicator that does effectively the same thing. By the sounds of it, this would be out of your price range, but it would do the job (we have a 15TB data centre mirrored using EMC's SRDF and another one using Volume Replicator).
In terms of free ways to do it, it will really depend on how sync'd the two offices need to be. If it's instantaneous, then you will need to have one master server and both sites pointing to it. Others have mentioned AFS, but that is also non trivial. If the synch doesn't have to be instantaneous, then perhaps a regular rsync tunneled through SSH would do the trick. CVS may also help, depending on the data you have.
Agreed. The best presenters are well versed in their subject area and are confident enough to not only put that across to the audience, but handle tricky questions while in mid presentation. I've always been impressed by folks who can stop what their saying, answer a difficult question, then pick up where they left off.
The best recent example I can think of was a guy from NetAPP who basically didn't have anything prepared and just stood up and talked for 45 minutes. That sounds boring on the face of it, but I came away knowing a lot more about their technology than before and actually enjoyed listening to him. He clearly knew what he was talking about and this came across in how passionate he was when speaking about it.
I don't think it's all Powerpoint's fault. If you don't have a clue how to present to people then even the best tool isn't going to help. A lot of people whom I see present need to take some kind of course or at least practice amongst peers before presenting in an important meeting. This is what I did to try and hone my skills as I found myself presenting more often. I still wouln't regard myself as an expert by any means, but at least I can be understood by the audience most of the time.
Some things that I have found to be effective over time have been whiteboard/flipcharts and handouts. On a whiteboard, I feel I'm much better equiped to handle the ad-hoc question, or explain a concept more fully. Too many presenters paint themselves into the slideware corner, by not having any other visual aids. The best lecturers I had at University used the whiteboard as much as the overheads/PPT slides. The other thing I do when giving a presentation, particularly if it's rather dense, is to provide a handout, either in the form of a white paper or a summary of the presentation with the speakers notes. In fact, most of my presentations are actually a summary of the technical paper I wrote in the first place. A PDF or printed copy is much easier for the audience to digest later, when they've forgotten everything I just talked about:) The other big factor is the presenter voice. If you just drone on and on, people are going to switch off in the first minute or two.
All of that said, I have major issues with PPT on a usability and document portability front. A lot of Tufte's ideas are valid in my view and at least some of them I hope to incorporate into my work...
Good point. Actually installing the patches is a piece of cake, but I agree that distribution isn't as good as other vendors. We usually end up scripting it and patching from NFS shares.
Re:How does this help us, or Sun
on
Solaris 9 x86 Review
·
· Score: 5, Insightful
I often ask myself (or am asked by someone else) the question "why do you use SPARC/Solaris when there are supposedly cheaper alternatives?"
- Reliability: I have found Solaris to be the most stable and reliable server OS I've used (obviously IBM Mainframes and VMS are even more so, but I haven't really got a lot of experience with them). This counts for a lot, as most of the sites I've looked after demand this in their requirements. Linux is pretty good too, but I have found Solaris to be absolutely rock solid.
- Relatively secure. It doesn't come out of the box with all the security you'd want, but its pretty damn easy to automate the install to get it to be so. It's certainly more secure than Windows:)
- Support. I've always found Sun support to be excellent, either through SunSolve or the call centres. This is a big plus.
- Application and Hardware support. Most of the big software and hardware vendors have a Solaris version. Makes it a bit easier to deploy things when you know they are certified.
- Ease of admin. I have become so used to Solaris that sys admin is pretty much second nature (but you could easily say the same for other OS's I guess).
- Standards compliant. All of the usual suspects are here: NFS, NIS, LDAP etc etc. Some of these standards were even developed by or with Sun's help. Interoperability is easier with Solaris than manyh other OS's IMHO.
I could go on and on, but that's my 2c. Obviously other folks will be able to say the same things for other OS's, but Sun have earned my trust (and dollars) over the years. I don't see any reason to fix what is already working, particularly with the price of their hardware coming down dramatically.
Re:How does this help us, or Sun
on
Solaris 9 x86 Review
·
· Score: 3, Interesting
Agreed, that is a good reason, but it gets down to the availability of apps to go with it. In the Oil and Gas world, the big players, like SLB, Paradigm and Landmark (to name but a few) have suites that run on SPARC/Solaris, IRIX and Linux. It would nice to have those apps on Sol x86, but the market probably isn't there.
Those Blade2500's are pretty nice machines though, we have a couple internally.
Re:How does this help us, or Sun
on
Solaris 9 x86 Review
·
· Score: 3, Informative
It rather does beg the question, doesn't it? I run mostly SPARC/Solaris systems with some SGI's and Linux mixed in and I can't think of many situations where I'd go for the x86 version. If I wanted an *all* Solaris shop, there'd be a case for it on low end boxes, but that's unlikely as monoculture of any type is bad(tm):) Besides, second hand SPARC systems can be had cheaply on eBay or even new from Sun/Fujitsu these days.
I wasn't surprised that Sun dropped Sol x86, but apparently there was a revolt from the market. It was a bit of a shock to me really. I wouldn't worry about Sun's R&E resources, that is one thing they seem to have consistently funded over the years.
It's a little unclear if you are looking at an outsourced backup service (i.e. backing up to a service company over the network) or a disk based backup system.
I will make the assumption that you are looking at a disk based system. There are a number of factors you need to look at:
1, Capacity required. That is, how much space will you need now and in the future? Tape systems offer advantages in being (generally) able to handle large amounts of data, particularly if it's compressable.
2, Restore time. This is very important. How quickly do you need to get data back? A disk based system will obviously have advantages in that it will give you near instantaneous restores. Tape can also provide fast restores, if you get sufficiently expensive drives (e.g. AIT-3, 9840b/c).
3, Retention period. This is related to capacity requirements, particularly in disk based systems. How far back do you want to go? Tape offers the ability to easily store media offsite and import it back in.
4, Budget. Notice that I put this last. The other considerations IMHO are more important than price, particularly if you really value your data. Spend the money and get a quality system, regardless of whether you get disk or tape.
Without knowing a lot about your situation, here are some options to think about:
- If restore time is key, look at getting a NAS (Network Attached Storage) or ATA array and use rsync or standard utility (e.g. tar, dump etc) to backup to the array. You can then use a commercial product like NetBackup or Networker or an open source utility like AMANDA to archive the data to tape if you want more control over versioning and need some offsite DR.
- If capacity is your goal, tape is the most cost effective option. In spite of appearances, high capacity tapes like the 9840B, AIT-3 and LTO2 will often give lower overall cost per megabyte than disk. You will often also get more scalability, as the libraries are easily expandable in terms of slots and/or drives. You can multiplex backup/restores over several drives to get you the pwerformance you're after.
- Use snapshots. NetAPP and other NAS vendors offer this and they backup only what changes. They also give you very quick recovery times.
I favour a hybrid disk and tape approach. That is, backup to a cheap disk array using the method of your choice and then backup the array to tape (full backup on weekends, incrementals on weekdays). Use a four week cycle for your tapes, with a monthly offiste rotation. If you have a bit of budget, then HSM software like SAM-FS or Veritas Storage Migrator make the job even easier. They are basically filesystems that use a disk cache and backup/restore to/from tape in the background. You can then NFS/SAMBA mount the filesystem on any machine. You can also write multiple copies of the same files, which is handy.
Solaris has ACL's and also Role Based Access Control (RBAC), just like Windows and has had for some time. It has a host of other nice admin features which would be off topic to go into here.
Slashdot Mirror
This looks pretty good for the network performance alone, plus the RAID controller is pretty nifty. If it works nicely under Linux (and nVidia's site appears to have drivers) it'll make a good chipset for a low end server. I like what I've seen of the new AMD stuff, particularly their new system design (Hypertransport etc) . Had a presentation from Sun recently on their coming Opteron lineup and that looks pretty sweet too (looks like IBM is offering them as well).
Agreed. I tend to run Sun equipment because it's rock bloody solid and just works. The other reason is Solaris. I find it to be among the more reliable and easy to admin server O/S's out there. I just wish DEC was still around. Loved those Alpha systems (yes, yes I know HP sorta still has 'em, but Compaq effectively killed the platform years ago). If that makes me behind the times or whatever, I couldn't care less -- I'll take my uptime, scalability and reliability thanks. Would love to have a G5 on my desk though :)
Patching Solaris really isn't that hard. You either download the patch you need and install it or simply get the whole recommended cluster and run the built in script. I find it no more difficult than any other platform. If there are pre-requisite patches, then the web page for the patch on Sunsolve tells you what they are. Pretty simple if you ask me. Granted, OSX is pretty nice too and patching my iBook is fairly trivial.
As to your other point, I have rarely seen Solaris patches outright break stuff to the point of rendering applications unusable. I normally feel pretty confident that they'll work as advertised.
I would argue that studio costs have come down significantly in the last 10 years, since DAW's (Digital Audio Workstations) came on the market. It is now possible to build up a semi pro recording rig for only a few thousand dollars now, whereas before the analogue multitrack tape machines were often tens of thousands of dollars by themselves. Leaving aside the analogue vs digital debate (pointless IMHO), I think it's possible to produce very good recordings fairly cheaply with a PC/MAC based solution these days.
:)
The thing to remember is that the recording software is only a drop in the bucket compared to the cost of the other equipment. Good mics and outboard equipment is where the majority of the money goes (not to mention preamps and mixing desks if your needs extend that far).
Of course, a good pair ears and some skill is the most valuable asset in a studio
Sun has had the ability to do multiple system images on the same box for a while, but they've always been hardware partitioning only. The 4800/6800/12k/15k allowed you to run different domains on the same system, so long as you had the right combo of CPU and I/O boards. This was great if you had one of those systems, but not so hot it you had a workgroup level system (e.g. E450 or V880). I'm glad to see they've put software partitioning in the O/S so I can take a mid range system and chop it up into separate pieces. AIX and HP-UX have been able to do the software side thing for a while (but not the dedicated hardware piece, I believe).
This will help with consolidation and utilisation on existing machines, I think.
I just finished reading this book and agree with the recommendation. It tells the story of what happened at that remarkable facility over a period of about 10 years. It was certainly rather illuminating. A must, I think, for anybody interested in the history of computing.
The odd thing for me was how advanced systems like the Alto were and that so many of the innovations it contained within took so long to come to market.
Didn't say anything about Sun in particular. There aren't many outfits that do massive scalability like SGI, and those new Altix servers are very nice (we have a few ourselves). They aren't the only solution on the market and neither are SPARC's. BTW, solaris scales very niceley on big iron, just look at it on the top end FujitsuSeimens systems.
I think you rather proved my point that shared memory systems still have great performance characteristics.
Ahh, it depends!! Some apps require the sub processes to be able to communicate with each other, preferably over a very high speed interconnect. This is where NUMA and SMP are handy. They allow near linear scaling of performance as you add processors. This is not always the case with clusters. I can take most multithreaded apps and get predictable scalability on an SMP system, but not always on a cluster.The other thing to consider is if you need a single memory image, versus fragmenting over many nodes. Do you need cache coherency to go with it?
To give an example, the company I work for writes a piece of software that does reservoir simulation. It is parallel capable and can be run on cluster or SMP systems. An IBM p690 will run the simulation around double the speed of a PC cluster (using Myrinet and Xeons at nearly double the clock speed) with the same number of total processors. This tells me that for this particular application, I would strongly consider using SMP iron.
When it comes to databases, you have to look at how the engine is clustered. Is it "shared all" (where the nodes in a cluster access the entire database with shared storage) or "shared nothing" (where the database is broken down into chunks with each node having its own piece). Last time I checked, Oracle was shared all and DB/2 and SQLSERVER were shared nothing. This makes a huge difference in how you set them up.
The other factor when it comes to SMP vs Clustering, is maintenace and management of the nodes. Even very well run clusters require more admin time than single systems.
Rubbish.
Two good friends of mine are working in Mumbai for the same company I work for. They were able to obtain working visas without a problem.
It's actually a good sign when upper management wants to talk. You really can take the opportunity to work through issues that affect the department and the business. The best IT groups form a partnership with their user base and use it to work on issues together.
My advice:
1, Confront the real issues. On both sides. There is no point in trying to dance around major problems if they are affecting the organisation.
2, Be honest. The old cliche about it being the best policy is actually pretty true. It is the best policy. Face the brutal truth and it sets you free in a funny kinda way. Once you do this, you can set about fixing the problem.
3, Be tactful. Ranting, raving and whining get you nowhere, particularly with management. I have found that airing concerns, but couched in a way that show the impact on the organisation as a whole works well. Upper level managers can relate to it that way.
4, Actually show concern about the direction of the business. This goes along with the whole partnership deal. It engenders a bit of trust, which can be a huge help to you later when you ask them to take a chance on a big risky project you want money/time/people for.
5, Listen to what they have to say. This is obvious, I know, but it's amazing how much you learn when actively listen to where their pain is. They'll love you forever if you can show exactly how to ease it.
6, Have a game plan. Think of what you want to talk about first, even if it's in broad terms. You don't waste people's time that way.
I would treat this as a golden opportunity to make progress on issues and generally gain a better understanding of their business and them of yours. You may never get a better chance, particularly in an informal situation like dinner. If at the end of the day nothing changes, you at least gave it a shot.
Nothing has really changed in terms of building a successful business, IMHO. Providing a product/service that your customers want to buy at a price they are prepared to pay and keeping them happy will increase your chances of success. Of course, there's more to it than that, but in effect that's it.
A book I found very useful is "Good to Great" by Jim Collins, which covers why some companies fall apart and some succeed. It came down to good leadership, top quality people, understanding of core business, discipline and facing the hard realities. Worth a read for anyone involved in private enterprise.
All capitalism does is allow people to act in their own rational self interest, which they would do anyway as IMHO that is a fundamental human trait. At under a capitalist system, they can be compensated for their own work and then have the freedom to spend the fruits of their labours in any way they choose (in the pure capitalist system, anyway). That doesn't sound like exploitation of people to me.
Now don't get me wrong, there are corporations and individuals who have done terrible things to both the environment and other people in the name of maximising profits. Those involved should go to jail, not be the catalyst for some structural change in the economy. Most folks are decent and law abiding, so why should their private property rights be abolished because of the actions of a few individuals?
In any case, I would argue that Communism caused a great deal more damage to people's rights and the environment than capitalism. How many people were killed in the USSR and China in the last century? You can add religious fundamentalism to the list as well.
What you are describing is not quite libertarianism. For it to qualify, there would have to be some system to enforce contracts and preserve private property rights. Without a legislature and a court system, that cannot happen. What we see in places like Somalia is coercion and force used to resolve private disputes, which is the antithesis of a libertarian (IMO) point of view.
Pure any "-ism" can never work anyway, whether it's Libertarianism, Communism et al, as you're never going to get a country full of people to agree on the best way to run it. Better, IMHO, to have the representative system as in most countries, where people get to have a say periodically and elect those whom they think will most effectively cary out their wishes.
Agreed. Good on Sun for stepping up and providing an alternative.
I think the reason why Excel has become such a standard is that it's a bloody good spreadsheet in it's own right. I couldn't say the same for Word and Powerpoint (which have gotten harder to use and more bloated as time has gone on), but Excel beats other mainstream spreasheets I have used hands down. There are obviously more specialised products out there that do a great job in certain disciplines (such as SAS or Mathematica) but Excel is great for general purpose stuff like I do...
In my experience, those who earn those kinds of salaries are rarely in an easily replacable position. Furthermore, they are often given whichever tools give them the ability to do their jobs most efficiently. If that deviates slightly from a "standard" then so be it. The 150k a year salary might look high on the P&L, but if they are good they add substantially more value than that to the organisation (i.e in new sales, new ideas or increased efficiency). Their pay starts to look cheap then....
A very good point. I work in IT, but I actually really like what I do. Furthermore, I can't imagine not doing it (well, maybe unless I was a musician). I think it's important to have job you like, as you're at work anywhere from 8 - 12 hours a day. Added to that, if you enjoy it, you're generally better at it, get paid more because of your skills and probably be happier in other parts of your life ('cause you're not bitching about work all the time).
I am constantly amazed by folks who get into a career simply because it pays well, but are otherwise miserable. These same people often times will bitch about their work but not do anything about it. I saw this a lot when I worked in the Government sector. I guess they become so entrenched that they see no way out. Distressingly, I also see this in education from time to time, particularly high school teachers, but some times at tertiary level.
...then I would consider building a SAN with replication. High end storage solutions using HDS and/or EMC gear fix this problem by enabling remote block for block copy of data between identical arrays. Veritas also makes a product called Volume Replicator that does effectively the same thing. By the sounds of it, this would be out of your price range, but it would do the job (we have a 15TB data centre mirrored using EMC's SRDF and another one using Volume Replicator).
In terms of free ways to do it, it will really depend on how sync'd the two offices need to be. If it's instantaneous, then you will need to have one master server and both sites pointing to it. Others have mentioned AFS, but that is also non trivial. If the synch doesn't have to be instantaneous, then perhaps a regular rsync tunneled through SSH would do the trick. CVS may also help, depending on the data you have.
Agreed. The best presenters are well versed in their subject area and are confident enough to not only put that across to the audience, but handle tricky questions while in mid presentation. I've always been impressed by folks who can stop what their saying, answer a difficult question, then pick up where they left off.
The best recent example I can think of was a guy from NetAPP who basically didn't have anything prepared and just stood up and talked for 45 minutes. That sounds boring on the face of it, but I came away knowing a lot more about their technology than before and actually enjoyed listening to him. He clearly knew what he was talking about and this came across in how passionate he was when speaking about it.
I don't think it's all Powerpoint's fault. If you don't have a clue how to present to people then even the best tool isn't going to help. A lot of people whom I see present need to take some kind of course or at least practice amongst peers before presenting in an important meeting. This is what I did to try and hone my skills as I found myself presenting more often. I still wouln't regard myself as an expert by any means, but at least I can be understood by the audience most of the time.
:) The other big factor is the presenter voice. If you just drone on and on, people are going to switch off in the first minute or two.
Some things that I have found to be effective over time have been whiteboard/flipcharts and handouts. On a whiteboard, I feel I'm much better equiped to handle the ad-hoc question, or explain a concept more fully. Too many presenters paint themselves into the slideware corner, by not having any other visual aids. The best lecturers I had at University used the whiteboard as much as the overheads/PPT slides. The other thing I do when giving a presentation, particularly if it's rather dense, is to provide a handout, either in the form of a white paper or a summary of the presentation with the speakers notes. In fact, most of my presentations are actually a summary of the technical paper I wrote in the first place. A PDF or printed copy is much easier for the audience to digest later, when they've forgotten everything I just talked about
All of that said, I have major issues with PPT on a usability and document portability front. A lot of Tufte's ideas are valid in my view and at least some of them I hope to incorporate into my work...
Good point. Actually installing the patches is a piece of cake, but I agree that distribution isn't as good as other vendors. We usually end up scripting it and patching from NFS shares.
I often ask myself (or am asked by someone else) the question "why do you use SPARC/Solaris when there are supposedly cheaper alternatives?"
:)
- Reliability: I have found Solaris to be the most stable and reliable server OS I've used (obviously IBM Mainframes and VMS are even more so, but I haven't really got a lot of experience with them). This counts for a lot, as most of the sites I've looked after demand this in their requirements. Linux is pretty good too, but I have found Solaris to be absolutely rock solid.
- Relatively secure. It doesn't come out of the box with all the security you'd want, but its pretty damn easy to automate the install to get it to be so. It's certainly more secure than Windows
- Support. I've always found Sun support to be excellent, either through SunSolve or the call centres. This is a big plus.
- Application and Hardware support. Most of the big software and hardware vendors have a Solaris version. Makes it a bit easier to deploy things when you know they are certified.
- Ease of admin. I have become so used to Solaris that sys admin is pretty much second nature (but you could easily say the same for other OS's I guess).
- Standards compliant. All of the usual suspects are here: NFS, NIS, LDAP etc etc. Some of these standards were even developed by or with Sun's help. Interoperability is easier with Solaris than manyh other OS's IMHO.
I could go on and on, but that's my 2c. Obviously other folks will be able to say the same things for other OS's, but Sun have earned my trust (and dollars) over the years. I don't see any reason to fix what is already working, particularly with the price of their hardware coming down dramatically.
Agreed, that is a good reason, but it gets down to the availability of apps to go with it. In the Oil and Gas world, the big players, like SLB, Paradigm and Landmark (to name but a few) have suites that run on SPARC/Solaris, IRIX and Linux. It would nice to have those apps on Sol x86, but the market probably isn't there.
Those Blade2500's are pretty nice machines though, we have a couple internally.
It rather does beg the question, doesn't it? I run mostly SPARC/Solaris systems with some SGI's and Linux mixed in and I can't think of many situations where I'd go for the x86 version. If I wanted an *all* Solaris shop, there'd be a case for it on low end boxes, but that's unlikely as monoculture of any type is bad(tm) :) Besides, second hand SPARC systems can be had cheaply on eBay or even new from Sun/Fujitsu these days.
I wasn't surprised that Sun dropped Sol x86, but apparently there was a revolt from the market. It was a bit of a shock to me really. I wouldn't worry about Sun's R&E resources, that is one thing they seem to have consistently funded over the years.
It's a little unclear if you are looking at an outsourced backup service (i.e. backing up to a service company over the network) or a disk based backup system.
I will make the assumption that you are looking at a disk based system. There are a number of factors you need to look at:
1, Capacity required. That is, how much space will you need now and in the future? Tape systems offer advantages in being (generally) able to handle large amounts of data, particularly if it's compressable.
2, Restore time. This is very important. How quickly do you need to get data back? A disk based system will obviously have advantages in that it will give you near instantaneous restores. Tape can also provide fast restores, if you get sufficiently expensive drives (e.g. AIT-3, 9840b/c).
3, Retention period. This is related to capacity requirements, particularly in disk based systems. How far back do you want to go? Tape offers the ability to easily store media offsite and import it back in.
4, Budget. Notice that I put this last. The other considerations IMHO are more important than price, particularly if you really value your data. Spend the money and get a quality system, regardless of whether you get disk or tape.
Without knowing a lot about your situation, here are some options to think about:
- If restore time is key, look at getting a NAS (Network Attached Storage) or ATA array and use rsync or standard utility (e.g. tar, dump etc) to backup to the array. You can then use a commercial product like NetBackup or Networker or an open source utility like AMANDA to archive the data to tape if you want more control over versioning and need some offsite DR.
- If capacity is your goal, tape is the most cost effective option. In spite of appearances, high capacity tapes like the 9840B, AIT-3 and LTO2 will often give lower overall cost per megabyte than disk. You will often also get more scalability, as the libraries are easily expandable in terms of slots and/or drives. You can multiplex backup/restores over several drives to get you the pwerformance you're after.
- Use snapshots. NetAPP and other NAS vendors offer this and they backup only what changes. They also give you very quick recovery times.
I favour a hybrid disk and tape approach. That is, backup to a cheap disk array using the method of your choice and then backup the array to tape (full backup on weekends, incrementals on weekdays). Use a four week cycle for your tapes, with a monthly offiste rotation. If you have a bit of budget, then HSM software like SAM-FS or Veritas Storage Migrator make the job even easier. They are basically filesystems that use a disk cache and backup/restore to/from tape in the background. You can then NFS/SAMBA mount the filesystem on any machine. You can also write multiple copies of the same files, which is handy.
Solaris has ACL's and also Role Based Access Control (RBAC), just like Windows and has had for some time. It has a host of other nice admin features which would be off topic to go into here.