Obviously you know very little about capital punishment in general, and the USA Patriot Act in particular.
In the US, there are basically two classes of capital crimes: murder and treason. Recent laws like the USA Patriot Act broadened the definition of both treason and murder to include those who aid and abet terrorists. But surely you don't suggest that spammers like Ralsky are hawking Viagra for Al Qaeda?
We have all the money, guns, and butter. And the clogged arteries to prove it. International business don't care about the millions of Eastern Europeans you include, because the vast majority can't afford a new pair of Nikes thanks to their Socialist predecessors. The US has the largest econonmy in the world by far, and even Euro corporations will not ignore that market. Politicians might try to fork the internet, but smart money will re-unify it in short order.
Microsoft Outlook Web Access, included with Exchange Server, is widely recognized to be the first real AJAX application. The 2000 version was the first browser app I every used that made me say "wow, how the hell did they do that?". No Java applet or ActiveX, but it felt like a real, usable desktop application. Context menus and everything, with few full-page refreshes.
Google has done quite a bit to elevate the profile of AJAX with the Slashdot crowd, but other people were definitely "really using it" long before Google.
It is very unlikely that any of the big major countries will go for a conventional war with each other, because they know that no one is going to emerge victorious.
I do recall Mr. Neville Chamberlain saying the much the same thing during the late 1930's. Never underestimate the stupidity of politicians.
P.S. Does a reference to Chamberlain violate Godwin's Law?
It was probably just the loading of all the client EXEs, DLLs,and template files from the server. Use FileMon to monitor all the I/.O requests your own PC makes when you fire up Word or Powerpoint. Now multiply that times the number of clients you had running MS Office from the file server.
According to TFA, you would never run more servers than CPUs in protection. That is utter bullcrap. ESX scales to 10 servers on a 2 way box according to VMware. I have a GSX box running on a 2 way box, and I have 6 production boxes using 25% of the CPU at any given time. That means I could scale to 15 with little trouble. In other words, this new scheme costs me more, a LOT more, than it did before.
Umm... it's not very smart to run SQL Server in a virtual machine environment. RDBMS are far to memory and I/O intensive for such an application. It would be much faster (and far cheaper from a licensing standpoint) to run multiple segregated instances of SQL Server on the same host OS.
Just because you can do something with VMware, doesn't mean you should...
Support != development. Oracle will effectively kill PeopleSoft and JD prodcut lines by stagnation.
Oracle's obvious ratioanly in buying PeopleSoft/JD was to get their large customer bases, and transition them to future versions of Oracle's applications. PeopleSoft and JD customers are effectively in maintenance mode.
Supporting three large, and mostly redundant application suites is not cost-effective from the vendor standpoint. My own company has experienced this "buyout and switch" practice for several of our major applications over the last few years. It's the nature of the software business.
Not unless you were running Citrix on the server. There has never been an "office server" for things like Word, Excel, and Powerpoint. (Unless you count a document managment system such as SharePoint Portal Server, which works with those apps but has none of their functionality).
That was just regular old office programs being loaded off a file server. There were still some files and a croapload of registry keys installed on the client machine, but the main EXEs and DLLs were loaded from the file server. So you could update the file server, and, in theory, everybody's copy of office would be upgraded. But the programs themselves still all ran on the client.
That's a far cry from a completely browser-based application that requires no plug-ins. Most folks love web-based email because they can get at it any time, from anywhere with a web browser. Go to Kinko's, borrow a friend's PC, etc. The trade-off in losing "rich client" functionality was more than worth the mobility people gained. GMail has added back a lot of that "rich client" experience with AJAX tricks, presumably GOffice will do the same.
Your story underscores the real point: hardware reliability is typically not the problem these days.
Software failures are (and this includes mySQL and PostGres) are far more common than hardware failures. Data corruption is a software failure, despite IBM's claim otherwise. (A database application should have a strong integrity-checking mechanism to ensure that a corrupt page is never written to a transaction log. If the data corruption occured in memory as they say, their log-writing process should have caught it and aborted the transaction. I have a feeling hardware was not the root cause of your failure, but I suppose corruption at the disk-driver level could make the failure undetectable to the DB application.)
Clustering, even shared-nothing clustering, cannot protect against failures of this sort when the clusters must share data in real-time. At least not without a whole lot of application-aware code to check the integrity of data before accepting it on the 2nd cluster node, which will slow the whole transaction down.
I have seen DB mirroring devices which act like a BIG-IP for database servers, sending the same transactions to two machines. But these introduce another bottleneck and another single point of failure, and presumably software problems on one DB server will also occur on the other servers when they are fed the same SQL data. And recovery is still a big problem.
Google does not have to worry about ACID compliance in their database. From what I've read about the google file system, cluster nodes lazily share new data amongst themselves. Serving up old data is explicitly allowed.
To cluster something like an OLTP database, every node has to be immediately informed about updates to the data, and they all have to report back that they have said data intact before the transaction commits. This can be something of a problem when you have hundreds of thousands of updates per second happening.
And of course, you need to have a method to rapidly bring back into sync a sever which has been out of commission for a while before it comes back online.
The only way I've seen to do that is to have some sort of high-speed shared interconnect between nodes, and some cluster-awareness in the application to handle synchronization. That is currenlty very expensive, especially if your some of your nodes are in California, and the rest are in Chciago.
Shared-nothing clusters simply require high-speed interconnects for transactional applications. Data changes must pushed everywhere, immediately, before the transaction commits. I don't see how you get around that.
Exchange 2003 SP1 raises that limit to 75 GB for the Standard Edition. I think they finally realized the 16 GB store limit, which seemed huge for a SMB just a few years ago, is rediculous now.
That said, the exchange information store does use single-instance-storage and compression. So that 16 GB limit actually translates to something like 50+ GB when compared with an mbox or maildir style system. Still, that's only enough for ~25 people now that everyone expects a 2 GB mailbox.
Of course the default with Windows is root, but hey....
Actually, in the corporate arena at least, this is completely false. By default, when a computer is joined to a windows domain, all domain accounts have only basic user privileges, not administrator or power user privileges. And it's been that way since Windows NT 4.0 circa 1996.
The real problems come from home & workgroup installations of Windows, where root is the default. And of course, so much poorly-written windows software requires admin or power user rights to run, that many companies punt assign all their users higher privileges.
How many people work in your company? Five? That's the only way I could see this working in real life.
And what is this about "executives private email"? Do they get a second private mailbox or something? So everything is *not* public, then, right? Do they have to remember to use this other address whenever they need to discuss personnel or financial issues? Relying on a busy executive to remember to switch email accounts when dealing with something sensitive is a recipie for disaster.
Congratsulations for perpetuating the stereotype of the dickhead IT guy without a clue about the way the real world works.
I cannot believe you were not fired for exposing all email archives to everyone in the firm. Didn't your HR people have something to say about your new policy? Does the CEO know he or she should not discuss personnel or M&A issues via email without using PGP or S/MIME?
The shaking-out period ended a long time ago. At least in Chicago, Illinois, USA.
I have lived in 5 different apartments and condos since 1998, and I've had multiple DSL and cable internet choices in all of them. I've used Earthlink, SBC, and Speakeasy DSL. As well as my current RCN cable. All providers worked with no real problems, and with ~2 week installation. I kept switching to get the "new subscriber" deals each time I moved.
My only problems were not really DSL-technology related. For about three months Eathlink's POP was totally over-subscribed (the 3rd hop router was the culprit). And other than Speakeasy, it took a lot of holding to get a real, live person on the phone to talk about billing or what-not. But DSL works, in this major metro at least.
Plus, there's actually some competition in the market. Brandbandreports.com lists 18 DSL (!) and 2 cable providers that service my address.
Why are "Buttfuck" and "Podunk" always reported as being in in Illinois or Indiana? I've lived in both of those states, and they're fairly cosmopolitan these days. I actually think "Buttfuck" is in upstate New York, and "Podunk" is just northwest of Sacramento, CA.
It is not illegal to be a monopoly or maintain a monopoly in the US. If you're so successful that you beat the pants off everyone else in your market, that's fine.
It is only illegal to use your monopoly position to shut competition out of the market through preferential supply, below-cost pricing, or back-room dealing.
For example, most public utilities, ILECs, Major Leage Baseball, and ICANN are all monopolies that are legal in the US.
In my tests on mixed binary and text data, 7-zip with LZMA and solid archiving came out on top in terms of compression ratio. It's very slow in this mode, but the best choice for archiving on Windows if you want as much compression as possible. It's also free and open source, and there is a Linux/BSD port available.
Hmmm... the Linux kernel was initially fast, stable, and built by a very small number of very smart people. However, it gets more and more unstable as new features are added and support for more hardware is added. The development process becomes unmanageable, because hundreds of developers are spread across the globe, speak different languages natively, and communitacate at irregular intervals. Each developer has pet features and subsystems they want to work on, while skimping on other "grunt work" tasks. Some silly mistakes are made. But backwards compatability must be maintained at all costs - we can't go around breaking applications from 10 years ago, no matter how stupid added experience makees those decades-old design decisions appear.
This sounds remarkably like the development history of another popular OS...
Right now there is some slashdot-reading marketing drone at MS thinking about how great C' sounds -- you know, because everyone knows that prime steaks are better than sharp cheese.
Slashdot Mirror
Obviously you know very little about capital punishment in general, and the USA Patriot Act in particular.
In the US, there are basically two classes of capital crimes: murder and treason. Recent laws like the USA Patriot Act broadened the definition of both treason and murder to include those who aid and abet terrorists. But surely you don't suggest that spammers like Ralsky are hawking Viagra for Al Qaeda?
We have all the money, guns, and butter. And the clogged arteries to prove it. International business don't care about the millions of Eastern Europeans you include, because the vast majority can't afford a new pair of Nikes thanks to their Socialist predecessors. The US has the largest econonmy in the world by far, and even Euro corporations will not ignore that market. Politicians might try to fork the internet, but smart money will re-unify it in short order.
Microsoft Outlook Web Access, included with Exchange Server, is widely recognized to be the first real AJAX application. The 2000 version was the first browser app I every used that made me say "wow, how the hell did they do that?". No Java applet or ActiveX, but it felt like a real, usable desktop application. Context menus and everything, with few full-page refreshes.
Google has done quite a bit to elevate the profile of AJAX with the Slashdot crowd, but other people were definitely "really using it" long before Google.
I do recall Mr. Neville Chamberlain saying the much the same thing during the late 1930's. Never underestimate the stupidity of politicians.
P.S. Does a reference to Chamberlain violate Godwin's Law?
It was probably just the loading of all the client EXEs, DLLs,and template files from the server. Use FileMon to monitor all the I/.O requests your own PC makes when you fire up Word or Powerpoint. Now multiply that times the number of clients you had running MS Office from the file server.
I would hate to see your resume.
In real life, spelling and grammar do actually matter.
Of course, Slashdot doesn't exactly qualify as "real life"...
Umm... it's not very smart to run SQL Server in a virtual machine environment. RDBMS are far to memory and I/O intensive for such an application. It would be much faster (and far cheaper from a licensing standpoint) to run multiple segregated instances of SQL Server on the same host OS.
Just because you can do something with VMware, doesn't mean you should...
Support != development. Oracle will effectively kill PeopleSoft and JD prodcut lines by stagnation.
Oracle's obvious ratioanly in buying PeopleSoft/JD was to get their large customer bases, and transition them to future versions of Oracle's applications. PeopleSoft and JD customers are effectively in maintenance mode.
Supporting three large, and mostly redundant application suites is not cost-effective from the vendor standpoint. My own company has experienced this "buyout and switch" practice for several of our major applications over the last few years. It's the nature of the software business.
Peoplesoft and J.D. Edwards don't count?
Not unless you were running Citrix on the server. There has never been an "office server" for things like Word, Excel, and Powerpoint. (Unless you count a document managment system such as SharePoint Portal Server, which works with those apps but has none of their functionality).
That was just regular old office programs being loaded off a file server. There were still some files and a croapload of registry keys installed on the client machine, but the main EXEs and DLLs were loaded from the file server. So you could update the file server, and, in theory, everybody's copy of office would be upgraded. But the programs themselves still all ran on the client.
That's a far cry from a completely browser-based application that requires no plug-ins. Most folks love web-based email because they can get at it any time, from anywhere with a web browser. Go to Kinko's, borrow a friend's PC, etc. The trade-off in losing "rich client" functionality was more than worth the mobility people gained. GMail has added back a lot of that "rich client" experience with AJAX tricks, presumably GOffice will do the same.
Your story underscores the real point: hardware reliability is typically not the problem these days.
Software failures are (and this includes mySQL and PostGres) are far more common than hardware failures. Data corruption is a software failure, despite IBM's claim otherwise. (A database application should have a strong integrity-checking mechanism to ensure that a corrupt page is never written to a transaction log. If the data corruption occured in memory as they say, their log-writing process should have caught it and aborted the transaction. I have a feeling hardware was not the root cause of your failure, but I suppose corruption at the disk-driver level could make the failure undetectable to the DB application.)
Clustering, even shared-nothing clustering, cannot protect against failures of this sort when the clusters must share data in real-time. At least not without a whole lot of application-aware code to check the integrity of data before accepting it on the 2nd cluster node, which will slow the whole transaction down.
I have seen DB mirroring devices which act like a BIG-IP for database servers, sending the same transactions to two machines. But these introduce another bottleneck and another single point of failure, and presumably software problems on one DB server will also occur on the other servers when they are fed the same SQL data. And recovery is still a big problem.
Google does not have to worry about ACID compliance in their database. From what I've read about the google file system, cluster nodes lazily share new data amongst themselves. Serving up old data is explicitly allowed.
To cluster something like an OLTP database, every node has to be immediately informed about updates to the data, and they all have to report back that they have said data intact before the transaction commits. This can be something of a problem when you have hundreds of thousands of updates per second happening.
And of course, you need to have a method to rapidly bring back into sync a sever which has been out of commission for a while before it comes back online.
The only way I've seen to do that is to have some sort of high-speed shared interconnect between nodes, and some cluster-awareness in the application to handle synchronization. That is currenlty very expensive, especially if your some of your nodes are in California, and the rest are in Chciago.
Shared-nothing clusters simply require high-speed interconnects for transactional applications. Data changes must pushed everywhere, immediately, before the transaction commits. I don't see how you get around that.
Exchange 2003 SP1 raises that limit to 75 GB for the Standard Edition. I think they finally realized the 16 GB store limit, which seemed huge for a SMB just a few years ago, is rediculous now.
That said, the exchange information store does use single-instance-storage and compression. So that 16 GB limit actually translates to something like 50+ GB when compared with an mbox or maildir style system. Still, that's only enough for ~25 people now that everyone expects a 2 GB mailbox.
Microsoft Outlook has had this feature since the original realease in 1996. View->By Conversation Topic. Is Outlook "mainstream" enough for you?
Actually, in the corporate arena at least, this is completely false. By default, when a computer is joined to a windows domain, all domain accounts have only basic user privileges, not administrator or power user privileges. And it's been that way since Windows NT 4.0 circa 1996.
The real problems come from home & workgroup installations of Windows, where root is the default. And of course, so much poorly-written windows software requires admin or power user rights to run, that many companies punt assign all their users higher privileges.
You discovered booze and drugs and sex, eh?
How many people work in your company? Five? That's the only way I could see this working in real life.
And what is this about "executives private email"? Do they get a second private mailbox or something? So everything is *not* public, then, right? Do they have to remember to use this other address whenever they need to discuss personnel or financial issues? Relying on a busy executive to remember to switch email accounts when dealing with something sensitive is a recipie for disaster.
Finally, of course, your scheme violates a cardinal security rule, the principle of least privelege.
Do you do your network backups by uploading everything to Gnutella or eDonkey?
Congratsulations for perpetuating the stereotype of the dickhead IT guy without a clue about the way the real world works.
I cannot believe you were not fired for exposing all email archives to everyone in the firm. Didn't your HR people have something to say about your new policy? Does the CEO know he or she should not discuss personnel or M&A issues via email without using PGP or S/MIME?
The shaking-out period ended a long time ago. At least in Chicago, Illinois, USA.
I have lived in 5 different apartments and condos since 1998, and I've had multiple DSL and cable internet choices in all of them. I've used Earthlink, SBC, and Speakeasy DSL. As well as my current RCN cable. All providers worked with no real problems, and with ~2 week installation. I kept switching to get the "new subscriber" deals each time I moved.
My only problems were not really DSL-technology related. For about three months Eathlink's POP was totally over-subscribed (the 3rd hop router was the culprit). And other than Speakeasy, it took a lot of holding to get a real, live person on the phone to talk about billing or what-not. But DSL works, in this major metro at least.
Plus, there's actually some competition in the market. Brandbandreports.com lists 18 DSL (!) and 2 cable providers that service my address.
the jurisdiction of Buttfuck, Illinois
Why are "Buttfuck" and "Podunk" always reported as being in in Illinois or Indiana? I've lived in both of those states, and they're fairly cosmopolitan these days. I actually think "Buttfuck" is in upstate New York, and "Podunk" is just northwest of Sacramento, CA.
It is not illegal to be a monopoly or maintain a monopoly in the US. If you're so successful that you beat the pants off everyone else in your market, that's fine.
It is only illegal to use your monopoly position to shut competition out of the market through preferential supply, below-cost pricing, or back-room dealing.
For example, most public utilities, ILECs, Major Leage Baseball, and ICANN are all monopolies that are legal in the US.
In my tests on mixed binary and text data, 7-zip with LZMA and solid archiving came out on top in terms of compression ratio. It's very slow in this mode, but the best choice for archiving on Windows if you want as much compression as possible. It's also free and open source, and there is a Linux/BSD port available.
Hmmm... the Linux kernel was initially fast, stable, and built by a very small number of very smart people. However, it gets more and more unstable as new features are added and support for more hardware is added. The development process becomes unmanageable, because hundreds of developers are spread across the globe, speak different languages natively, and communitacate at irregular intervals. Each developer has pet features and subsystems they want to work on, while skimping on other "grunt work" tasks. Some silly mistakes are made. But backwards compatability must be maintained at all costs - we can't go around breaking applications from 10 years ago, no matter how stupid added experience makees those decades-old design decisions appear.
This sounds remarkably like the development history of another popular OS...
Oh, no. Now you've done it.
Right now there is some slashdot-reading marketing drone at MS thinking about how great C' sounds -- you know, because everyone knows that prime steaks are better than sharp cheese.