I'm an independent IT provider and I can't tell you how to find me. You have to know someone I currently work for, and be refereed over. The key is someone that fits personality wise with your company. Have you made you issues clear and with the right people? Have you made your frustrations known or is there just a lack of communication?
Let's flip this the other direction, and see it from the IT perspective? What's your monthly IT spend like? Are you paying for services or work and not receiving it? Is there a list of IT projects or tasks that are waiting on money? As an IT provider, my focus is on my customers, but the next thing on my mind is making sure my monthly bills are covered. Bill rates reflect a lot of things, everything from market size, and cost of living, down to if I have a day job and you're second fiddle for that 11am meltdown. Paying for a MSP styled package should give you a static costing for IT, which beats the snot out of an unexpected bill for a $1,000 because a system took a crap.
There are tools out there that an IT vendor or MSP can use to make life easier - things like a RMM package, but those come with a per-pc cost every month. RMM tools can make things like controling Windows 10 updates from happening, or at least not happening before you want. It also helps cut down on trip charges & response times if I can quickly see what you're dealing with and resolve it.
As to finding another IT vendor, chat with other business owners and see who they use. A referral generally is going be be better than opening the yellow pages or Google. Know that when you switch vendors there is going to be a large expense - you're paying the old firm to transition over knowledge and documentation to the new firm, plus the new firm will want to resolve any immediate issues in their eyes. There is many different ways to do the same thing in technology, and determining right and wrong is really a matter of opinion - which we all have.
Do not make business decisions based on a 3 year warranty - it generally only covers the hardware and returning the system back to square one. Make sure you have a solid backup plan, and remember that unless you're an emergency room, no one is going to die.
https://www.backblaze.com/b2/i... - Use an application that works, and you're set. If you want to be more cost aware, doing a local NAS and sync'ing what matters up to B2 centrally allows for more instant restores locally, but if the worst of events happens, you can pull the offsite data.
Seriously, this user has submitted many open ended, or inexperienced questions. Check out the users submission history: https://slashdot.org/~dryriver
No, all software anywhere will never happen. Period. You would have to force everyone to code in specific cross platform languages and frameworks, and force hardware vendors to only develop based around specific platforms and restrictions. It will never happen. Imagine not having Arduino as a platform, because it's limited power and function is great at embedded systems. Platforms are selected for specific software based on what the developer wants to create. Low power? Graphics power? CPU power? Different use cases need different platforms.
So having to deal with this with my GF - there is a legit need. For example, a dev will say 'do you want a responsive or non-responsive website'. Devs understand what responsive stands for - the accountant is looking for a 101 type answer. When you start talking about technologies, knowing what a wireframe is, and how it applies to a time line is critical. There is geek jargon for every discipline, and knowing a bit about what phrases mean to a non-tech is a legit concern. For example, if they're proposing a Flash based site, and the target mobile users, it's not a good fit.
Unless the dev is of the right mindset and patient, the customer can become really frustrated, to the point they're blindly accepting the dev's recommendations, and then there are issues when the finished product doesn't work.
There is a huge difference. RTOs no longer allow for the time to restore from it, but long term archival to take is very viable. It's power is that the medium is separate from the mechanics of the heads and drive motors. You can replace a drive with a newer one and read a few generations older tape - but if a hard drive has an internal breakdown, you're sending it in to a data recovery company.
So wrong in so many ways. Any reason you wouldn't purchase a 100 year certificate and just roll with it? Too bad about 1/3 of all Azure disk space is used for endpoint backup.
This reminds me of the leap-year calculating bug - Feb 29 2012, you couldn't generate a site because the default is to generate a certificate for 1 year, and well, Feb 29 2013 just doesn't exist.
http://blogs.msdn.com/b/windowsazure/archive/2012/03/09/summary-of-windows-azure-service-disruption-on-feb-29th-2012.aspx
They would spin off the Entertainment Division into a separate company to existing shareholders, but the risk would be that the exodus of MSFT stockholders after that. XBox is a huge brand that has lots of value that is looking up, not down. It may be a job saving tactic for Ballmer, but I doubt it would work.
Console is a VNC session - so you just need to connect to the port in the correct way. I know some folks who do it already, so it's got to be written somewhere.
Most bags are great, the big deal will be if you're doing a pro body (or standard body + grip) or a standard body. Lots of options that will depend on if you're hiking or hitting airports. I had to go the Lowepro route due to the pro body - and it'll fit a 70-200/2.8 and 3-4 lenses.
- For HyperV & Xen you need VT-enabled hardware, latest Intel & AMD only - Go for more than 4gb of RAM, part of this is about pushing a few limits, and you'll want to run a few VM's at the same time - Get a hardware RAID card - I think the LSI MegaRAID SATA-150 is about as standardized and supported as it gets - ESX/ESXi work fine, and Xen and HyperV should all work - Use laptop drives for your RAID set - they fit much nicer into cases (4x320's gives you 900gb) - Go for 2-3 network adapters - Intel or Broadcom only (10/100 is fine)
Pick a base OS and run VMWare Server - trust me on this. Instead of reinstalling the OS off of cd, you're mounting the cd in the VM and doing your installs without the legwork. You can also download pre-built demo appliances so you spend your time dealing with the product (Oracle, IBM, etc), rather than tweaking out your CentOS config.
If you're going to stick with VMWare ESX/ESXi you can get any server hardware from the last 3 years. Sun x4100's, Dell 1750/2650's, HP DL380/385's all work fine, though RAM is mostly still expensive:-p.
So when you're purchasing power from the grid and you're metered not on use, but on peak draw, this will save you a LOT of money. Coordinating the power on of a number of systems which draw a lot at power on verses their normal draw (think turning on 100 laser printers all at the same time!).
Now that this has happened, how do we secure BGP!?
on
Pakistan Blocks YouTube
·
· Score: 2, Interesting
So, what are the steps to secure the IP space in BGP? How likely is this to occur again? What if Pakistan couldn't subnet accurately?
sBGP, where are you?
So there are huge swaths of IP space tied up in entities which don't need any where near as many as before NAT. If ARIN's requirements for usage were enforced then we may be fine for the next 10 years. Anyone with a Class A needs to figure out what they're doing and return some major swaths of IP space:
1.0.0.0/8 - IANA 2.0.0.0/8 - IANA 3.0.0.0/8 - GE 4.0.0.0/8 - Level 3 5.0.0.0/8 - IANA 6.0.0.0/8 - DoD 7.0.0.0/8 - DoD 8.0.0.0/8 - Level 3 9.0.0.0/8 - IBM 10.0.0./8 - NAT (we all love it) 11.0.0.0/8 - DoD
Now as an ISP, I want to be multi-homed, but the only "legit" mannor to do this is via an IP allocation from ARIN, otherwise you'll be forcing a renumber on your clients (not a happy thing).
So here is what people need to realize, that the "Internet" is only as good as the computers connected to it. Why would you want to connect to a network - easy, the content you can access. Same as for the office, for the home, and for the internet. You connect to a network to access content. You can create a network, like many a community wifi-based networks have built, but unless you connect to something with content, no one else will want to connect to your network.
So the issue with network neturality is that the provider of content has to exist, otherwise, the client of the content (consumers, etc) will never connect. Content (news, video, shopping, anything served up) is why people connect to the network (internet). By charging consumers to connect to this network, a provider should be able to cover their costs, plus make a profit. There are different levels of providers of IP, and based on what level you're connecting at (10 mpbs mostly pull is going to cost you while an OC-48 can peer with most providers for free) will decide your price per meg.
These peering agreements are what make the internet work. If the servers of Ebay are connected to AT&T, and you have Earthlink as your DSL provider, you rely on the peering agreement between AT&T and Earthlink. Now Ebay could be charged for their internet connection, and they do pay at some level, but AT&T likes their traffic out to the internet (mostly push), due to their ability to pull from the internet for their customers (mostly pull). Peering agreements are two way agreements, so a good ISP is going to balance their push customers (slashdot, cnn, yahoo) with their pull customers (home internet connections, corporate networks).
Now to charge a premium for a certain customer to get access to specific content at a faster speed than otherwise available becomes a very tricky issue. Who would charge this fee, and who would count that fee on their corporate balance sheet? Would it be the ISP for the server, the back haul provider, or the ISP for the user? What about the peering points which allow for the internet to work? These are setup as non-profits and do not charge for exchanging traffic (but look at the levels of traffic the SeattleIX handles http://www.seattleix.net/agg.htm)
So here are a few things to remember, as to why IT / SysAdmin / NetAdmin / Desktop support people should be appreciated:
1) At 2am, I have to answer my phone - only job in the company where the expectations don't end when you leave the office. Don't believe me, call your CEO at 2am and not get fired.
2) Service windows are in the middle of the night, and we are expected to do our major work at that time. Taking down the email server during business hours is not acceptable to anyone. But we still have to be in the office from 8:30-5, so why don't you figure out how many hours of OT you work. Thats right, we're "exempt" so no matter how many hours it takes, we don't get paid.
3) IT cannot be a line budgeted item. If we need something, waiting until the next years budget to get it isn't an option - no one asks you to band-aide you TPS reports. Technology and requirements constantly change, and new tech comes out all the time. Never mind we don't get $$ per each new hire for expanding our server infrastructure.
4) Complain about a $150 per hour bill rate? Try figuring out the cost to a business for a server outage. How about a network outage? 100 people at $25 per hour salary = $2,500 per hour of outage. Which is cheaper? Pay my rate, you don't pay for my education time, purchases of new hardware or software, or endless hours like spending the weekend setting up a SAN at home.
5) Never mind about the home computer questions you come and ask us about - we're here to help all your technology needs. Yes, its not work related, but damned if you don't ask us during work hours. And we have to keep current and remember that you have a wireless network with XXXXXXX WPA password. No one else is expected to keep such detailed records on you.
6) Developers, come on - you can do your job from anywhere. As long as you come in for meetings, you may have a better environment at home, and you will be more productive when you don't have others walk up and distract you. IT types have to be there to hold a users hand, and plug in the mouse that got unplugged somehow.
When $ is no issue, a tier 1 colocation provider with their own services would be the best option. They've got big pipes, and will work with you to have the additional services needed. I'd go as far to say that you're going to want to have a failover script that they would follow in the event of site A going offline. You'd need redundant equipment, or use a DR firm for getting back up.
First off decide of you really need Citrix or not. There are a few things it does well, mostly on a management / printing basis. Take a look at some sort of SSL Presentation box (F5 Firepass / etc) to do your presentation. Using basic Termainal Services works fine for some situations.
Now Microsoft is allowing 4 free instances of their OS when you're running on Windows 2003 R3 Enterprise/Advanced, and using Virtual Server 2005 R2. I know it's a MS hot dog next to VMWare's Prime Rib, but when $$ matters there is compromise to be had.
I've used ESX for Win2003 std Terminal Server - due to the users each mapping 4 printers back each (yea Windows Server with 35 people connected, each bringing 4 printers - didn't work well). There's a check box in ESX for "Citrix Workload".
In a perfect situation, I'd use Citrix to publish applications. I'd create 1-3 VM's on each server for each application published (5 apps = 5-15 VM's per server). Use Citrix to balance the load across those servers (or an external appliance). This would allow for a fairly consistant load across the servers without any additional features. If you're in it for the money, create 2 VM's per task and use the new Vitual Infrastructure 3 DRS feature to allow automatic VMotion if a single server gets overloaded.
Something to think about, but remember using a Vitrual platform has so many advantages to strictly hardware I'd overlook the Citrix people saying "no". Rebuilding a server in 3 short mouse clicks is just too amazing.
Easy - Think SAN - Apple XServe RAID + DNFStorage
on
Best Server Storage Setup?
·
· Score: 5, Insightful
So having thought about this a lot, here's what I would do:
1) Run a FC SAN as the backend. This allows you to connect anything you want without wondering what future technology will allow for - ATAoE, iSCSI, ???
2) Love thy Apple. XServe RAID's are 3u, 7tb (raw) and $13,000 - get a bunch - each controller see's 7 disks, set them up as a RAID 0 and uplink the thing to a FC switch.
3) Use DNFStorage.com's SANGear 4002 / 6002 devices to RAID 5 across the XServe RAID 0 LUN's. Your data security then can tolerate half of an XServe RAID going offline. RAID 6 allows for an entire unit to become DOA. Make sure to have an online spare or two.
4) Repeat - but remember, just because you can create it, doesn't mean you can reasonibly back it up.
Now the stupid question - what are you trying to do that would require this much space when you don't have the budget to get a "tested, supported, enterprise" solution? Building things is fun, but at some point you need to back up and say, "Am I willing to risk my company on my solution". EMC, HDS, IBM, HP and other big vendors are willing to step up and make sure your solution works, runs and will not fail (see that video with the SAN array getting shot?)
Slashdot Mirror
I'm an independent IT provider and I can't tell you how to find me. You have to know someone I currently work for, and be refereed over. The key is someone that fits personality wise with your company. Have you made you issues clear and with the right people? Have you made your frustrations known or is there just a lack of communication?
Let's flip this the other direction, and see it from the IT perspective? What's your monthly IT spend like? Are you paying for services or work and not receiving it? Is there a list of IT projects or tasks that are waiting on money? As an IT provider, my focus is on my customers, but the next thing on my mind is making sure my monthly bills are covered. Bill rates reflect a lot of things, everything from market size, and cost of living, down to if I have a day job and you're second fiddle for that 11am meltdown. Paying for a MSP styled package should give you a static costing for IT, which beats the snot out of an unexpected bill for a $1,000 because a system took a crap.
There are tools out there that an IT vendor or MSP can use to make life easier - things like a RMM package, but those come with a per-pc cost every month. RMM tools can make things like controling Windows 10 updates from happening, or at least not happening before you want. It also helps cut down on trip charges & response times if I can quickly see what you're dealing with and resolve it.
As to finding another IT vendor, chat with other business owners and see who they use. A referral generally is going be be better than opening the yellow pages or Google. Know that when you switch vendors there is going to be a large expense - you're paying the old firm to transition over knowledge and documentation to the new firm, plus the new firm will want to resolve any immediate issues in their eyes. There is many different ways to do the same thing in technology, and determining right and wrong is really a matter of opinion - which we all have.
Do not make business decisions based on a 3 year warranty - it generally only covers the hardware and returning the system back to square one. Make sure you have a solid backup plan, and remember that unless you're an emergency room, no one is going to die.
https://www.backblaze.com/b2/i... - Use an application that works, and you're set. If you want to be more cost aware, doing a local NAS and sync'ing what matters up to B2 centrally allows for more instant restores locally, but if the worst of events happens, you can pull the offsite data.
https://www.backblaze.com/blog... has more info
Seriously, this user has submitted many open ended, or inexperienced questions. Check out the users submission history: https://slashdot.org/~dryriver
No, all software anywhere will never happen. Period. You would have to force everyone to code in specific cross platform languages and frameworks, and force hardware vendors to only develop based around specific platforms and restrictions. It will never happen. Imagine not having Arduino as a platform, because it's limited power and function is great at embedded systems. Platforms are selected for specific software based on what the developer wants to create. Low power? Graphics power? CPU power? Different use cases need different platforms.
It comes down to who can interact with it how. Are you doing HD or 4K for the monitors? http://www.brightsign.biz/digi...
So having to deal with this with my GF - there is a legit need. For example, a dev will say 'do you want a responsive or non-responsive website'. Devs understand what responsive stands for - the accountant is looking for a 101 type answer. When you start talking about technologies, knowing what a wireframe is, and how it applies to a time line is critical. There is geek jargon for every discipline, and knowing a bit about what phrases mean to a non-tech is a legit concern. For example, if they're proposing a Flash based site, and the target mobile users, it's not a good fit. Unless the dev is of the right mindset and patient, the customer can become really frustrated, to the point they're blindly accepting the dev's recommendations, and then there are issues when the finished product doesn't work.
There is a huge difference. RTOs no longer allow for the time to restore from it, but long term archival to take is very viable. It's power is that the medium is separate from the mechanics of the heads and drive motors. You can replace a drive with a newer one and read a few generations older tape - but if a hard drive has an internal breakdown, you're sending it in to a data recovery company.
So wrong in so many ways. Any reason you wouldn't purchase a 100 year certificate and just roll with it? Too bad about 1/3 of all Azure disk space is used for endpoint backup. This reminds me of the leap-year calculating bug - Feb 29 2012, you couldn't generate a site because the default is to generate a certificate for 1 year, and well, Feb 29 2013 just doesn't exist. http://blogs.msdn.com/b/windowsazure/archive/2012/03/09/summary-of-windows-azure-service-disruption-on-feb-29th-2012.aspx
They would spin off the Entertainment Division into a separate company to existing shareholders, but the risk would be that the exodus of MSFT stockholders after that. XBox is a huge brand that has lots of value that is looking up, not down. It may be a job saving tactic for Ballmer, but I doubt it would work.
http://www.scarecrow.com/ - it's just a matter if your local market will support it..
Unless you're looking to use the drives after the fact...
Console is a VNC session - so you just need to connect to the port in the correct way. I know some folks who do it already, so it's got to be written somewhere.
Most bags are great, the big deal will be if you're doing a pro body (or standard body + grip) or a standard body. Lots of options that will depend on if you're hiking or hitting airports. I had to go the Lowepro route due to the pro body - and it'll fit a 70-200/2.8 and 3-4 lenses.
So off the top of my head:
- For HyperV & Xen you need VT-enabled hardware, latest Intel & AMD only
- Go for more than 4gb of RAM, part of this is about pushing a few limits, and you'll want to run a few VM's at the same time
- Get a hardware RAID card - I think the LSI MegaRAID SATA-150 is about as standardized and supported as it gets - ESX/ESXi work fine, and Xen and HyperV should all work
- Use laptop drives for your RAID set - they fit much nicer into cases (4x320's gives you 900gb)
- Go for 2-3 network adapters - Intel or Broadcom only (10/100 is fine)
Pick a base OS and run VMWare Server - trust me on this. Instead of reinstalling the OS off of cd, you're mounting the cd in the VM and doing your installs without the legwork. You can also download pre-built demo appliances so you spend your time dealing with the product (Oracle, IBM, etc), rather than tweaking out your CentOS config.
If you're going to stick with VMWare ESX/ESXi you can get any server hardware from the last 3 years. Sun x4100's, Dell 1750/2650's, HP DL380/385's all work fine, though RAM is mostly still expensive :-p.
http://metrix.net/ubiquiti-powerstation-2-ps218v-p-109.html Get two, set them up, alignment LEDs are on the back. It doesn't get much easier.
So when you're purchasing power from the grid and you're metered not on use, but on peak draw, this will save you a LOT of money. Coordinating the power on of a number of systems which draw a lot at power on verses their normal draw (think turning on 100 laser printers all at the same time!).
So, what are the steps to secure the IP space in BGP? How likely is this to occur again? What if Pakistan couldn't subnet accurately? sBGP, where are you?
This is the point where the company needs to step up and be clear in everything. Otherwise they should just shutdown the data center in Chicago.
So there are huge swaths of IP space tied up in entities which don't need any where near as many as before NAT. If ARIN's requirements for usage were enforced then we may be fine for the next 10 years. Anyone with a Class A needs to figure out what they're doing and return some major swaths of IP space:
1.0.0.0/8 - IANA
2.0.0.0/8 - IANA
3.0.0.0/8 - GE
4.0.0.0/8 - Level 3
5.0.0.0/8 - IANA
6.0.0.0/8 - DoD
7.0.0.0/8 - DoD
8.0.0.0/8 - Level 3
9.0.0.0/8 - IBM
10.0.0./8 - NAT (we all love it)
11.0.0.0/8 - DoD
Come on people - if you're going to force usage on us, then force them on all http://www.arin.net/policy/nrpm.html#four33
Now as an ISP, I want to be multi-homed, but the only "legit" mannor to do this is via an IP allocation from ARIN, otherwise you'll be forcing a renumber on your clients (not a happy thing).
So here is what people need to realize, that the "Internet" is only as good as the computers connected to it. Why would you want to connect to a network - easy, the content you can access. Same as for the office, for the home, and for the internet. You connect to a network to access content. You can create a network, like many a community wifi-based networks have built, but unless you connect to something with content, no one else will want to connect to your network.
So the issue with network neturality is that the provider of content has to exist, otherwise, the client of the content (consumers, etc) will never connect. Content (news, video, shopping, anything served up) is why people connect to the network (internet). By charging consumers to connect to this network, a provider should be able to cover their costs, plus make a profit. There are different levels of providers of IP, and based on what level you're connecting at (10 mpbs mostly pull is going to cost you while an OC-48 can peer with most providers for free) will decide your price per meg.
These peering agreements are what make the internet work. If the servers of Ebay are connected to AT&T, and you have Earthlink as your DSL provider, you rely on the peering agreement between AT&T and Earthlink. Now Ebay could be charged for their internet connection, and they do pay at some level, but AT&T likes their traffic out to the internet (mostly push), due to their ability to pull from the internet for their customers (mostly pull). Peering agreements are two way agreements, so a good ISP is going to balance their push customers (slashdot, cnn, yahoo) with their pull customers (home internet connections, corporate networks).
Now to charge a premium for a certain customer to get access to specific content at a faster speed than otherwise available becomes a very tricky issue. Who would charge this fee, and who would count that fee on their corporate balance sheet? Would it be the ISP for the server, the back haul provider, or the ISP for the user? What about the peering points which allow for the internet to work? These are setup as non-profits and do not charge for exchanging traffic (but look at the levels of traffic the SeattleIX handles http://www.seattleix.net/agg.htm)
When will everyone else "get it"?
So here are a few things to remember, as to why IT / SysAdmin / NetAdmin / Desktop support people should be appreciated:
1) At 2am, I have to answer my phone - only job in the company where the expectations don't end when you leave the office. Don't believe me, call your CEO at 2am and not get fired.
2) Service windows are in the middle of the night, and we are expected to do our major work at that time. Taking down the email server during business hours is not acceptable to anyone. But we still have to be in the office from 8:30-5, so why don't you figure out how many hours of OT you work. Thats right, we're "exempt" so no matter how many hours it takes, we don't get paid.
3) IT cannot be a line budgeted item. If we need something, waiting until the next years budget to get it isn't an option - no one asks you to band-aide you TPS reports. Technology and requirements constantly change, and new tech comes out all the time. Never mind we don't get $$ per each new hire for expanding our server infrastructure.
4) Complain about a $150 per hour bill rate? Try figuring out the cost to a business for a server outage. How about a network outage? 100 people at $25 per hour salary = $2,500 per hour of outage. Which is cheaper? Pay my rate, you don't pay for my education time, purchases of new hardware or software, or endless hours like spending the weekend setting up a SAN at home.
5) Never mind about the home computer questions you come and ask us about - we're here to help all your technology needs. Yes, its not work related, but damned if you don't ask us during work hours. And we have to keep current and remember that you have a wireless network with XXXXXXX WPA password. No one else is expected to keep such detailed records on you.
6) Developers, come on - you can do your job from anywhere. As long as you come in for meetings, you may have a better environment at home, and you will be more productive when you don't have others walk up and distract you. IT types have to be there to hold a users hand, and plug in the mouse that got unplugged somehow.
You missed the point - there is prior art here - which would invalidate the patent.
Yea, gotta work on the title, but Sir Mix-a-lot has some tracks out under a license like this:
m l?tid=126&tid=141&tid=187&tid=188
http://slashdot.org/articles/04/01/10/2042228.sht
Looks like the service shut down though:
http://weedshare.com/
When $ is no issue, a tier 1 colocation provider with their own services would be the best option. They've got big pipes, and will work with you to have the additional services needed. I'd go as far to say that you're going to want to have a failover script that they would follow in the event of site A going offline. You'd need redundant equipment, or use a DR firm for getting back up.
First off decide of you really need Citrix or not. There are a few things it does well, mostly on a management / printing basis. Take a look at some sort of SSL Presentation box (F5 Firepass / etc) to do your presentation. Using basic Termainal Services works fine for some situations.
Now Microsoft is allowing 4 free instances of their OS when you're running on Windows 2003 R3 Enterprise/Advanced, and using Virtual Server 2005 R2. I know it's a MS hot dog next to VMWare's Prime Rib, but when $$ matters there is compromise to be had.
I've used ESX for Win2003 std Terminal Server - due to the users each mapping 4 printers back each (yea Windows Server with 35 people connected, each bringing 4 printers - didn't work well). There's a check box in ESX for "Citrix Workload".
In a perfect situation, I'd use Citrix to publish applications. I'd create 1-3 VM's on each server for each application published (5 apps = 5-15 VM's per server). Use Citrix to balance the load across those servers (or an external appliance). This would allow for a fairly consistant load across the servers without any additional features. If you're in it for the money, create 2 VM's per task and use the new Vitual Infrastructure 3 DRS feature to allow automatic VMotion if a single server gets overloaded.
Something to think about, but remember using a Vitrual platform has so many advantages to strictly hardware I'd overlook the Citrix people saying "no". Rebuilding a server in 3 short mouse clicks is just too amazing.
So having thought about this a lot, here's what I would do:
1) Run a FC SAN as the backend. This allows you to connect anything you want without wondering what future technology will allow for - ATAoE, iSCSI, ???
2) Love thy Apple. XServe RAID's are 3u, 7tb (raw) and $13,000 - get a bunch - each controller see's 7 disks, set them up as a RAID 0 and uplink the thing to a FC switch.
3) Use DNFStorage.com's SANGear 4002 / 6002 devices to RAID 5 across the XServe RAID 0 LUN's. Your data security then can tolerate half of an XServe RAID going offline. RAID 6 allows for an entire unit to become DOA. Make sure to have an online spare or two.
4) Repeat - but remember, just because you can create it, doesn't mean you can reasonibly back it up.
Now the stupid question - what are you trying to do that would require this much space when you don't have the budget to get a "tested, supported, enterprise" solution? Building things is fun, but at some point you need to back up and say, "Am I willing to risk my company on my solution". EMC, HDS, IBM, HP and other big vendors are willing to step up and make sure your solution works, runs and will not fail (see that video with the SAN array getting shot?)