The average rate of return for the NYSE over the last ~150 years has been 8% annually compounded, there is no other investment vehicle which compares. If you aren't playing the game you are losing. Heck after the crash in September 2008 I increased my 401k contribution by 50% because I will probably never see a return greater than the climb out of this recession. If I'm wrong it won't really matter because modern society will cease to function and so the wealth wouldn't do me any good in any investment except perhaps bullets =)
Yeah, the GP was talking about big time storage, to me that means SAN =) And the ram cache I was thinking of is battery backed for writes which system ram obviously can't do, and even for SSD's a bit of ram cache allows you to reorder writes which can greatly improve throughput and life.
With hard disk failure rates approaching or surpassing 50% within the first 6 months for some manufacturers and lots (regularly) I'd argue that we've reached the point where SSD system drives Make Sense.
Uh, quit buying cheap crap... My failure rate across 170 servers and two SAN arrays is less than 1.5% per year.
Actually the near to midterm future is tiered storage with fast ram cache first, then SSD's and finally big slow, cheap disks (or perhaps big fast moderate cost disk if you have a database too large to fit into SSD cache and your application can't stand the high latency of SATA). This is the design of the new Sun storage servers and also the design that Netapp uses (I'm sure there's others but those are the two big ones I'm aware of) and I think it's the way that makes the most sense since it most efficiently utilizes the expensive (on a $/GB) SSD resources.
Ha, downtime to swap out hardware?!? That's the biggest reason to go VMWare, swap out servers or storage with zero downtime (vmotion and svmotion for the win). Heck with Windows Enterprise I can hot add ram without downtime (CPU's too but most stuff won't take advantage without a reboot).
.5 to 3MW/hr? That's a pretty big facility, the entire 500 person HQ facility for my S&P 500 company including our datacenter only has 650kW/hr of generating capacity.
Dude, the Valdez was 11 MILLION gallons, at 50k barrels a day it will take ~ two months for this to be as big. I'd hope we can stop this is less than a month, let alone two. Now this will be bigger in that it will affect a larger area, and a more commercially significant area due to the amount of seafood produced in the region, but to make this sound like it's the end of the world is a bit overdone.
Like Proliant, MSA, EVA, etc? HP turned the Compaq products into some of the best selling products in the market and became the #1 computer manufacturer.
Gopher was already waning pretty heavily by 97, its heyday was probably 93-95 because by late 1995 you had Windows 95 and Navigator making the GUI web very approachable.
Why not lower the cost to upgrade to the first version that supported a modern distribution method? Might bring in more upgrade revenue and lower your and your customers operating costs.
Your right, Oracle, ADP, and a few other large vendors are incompetent, doesn't change the fact that in practice Java code is no more portable than well written c/c++ code.
Flash+Flashblock should maintain your sanity while not requiring you to load up IE, and with the new plugin container in 3.6.4 it doesn't even take down the browser if/when it crashes =) Oh and Flash 10.1 also significantly reduces CPU load by using GPU offloading if available.
Since upwards of half the cost of the system is installation I don't think a bankrupt SolarCity or their creditors is going to go around pulling panels off your roof. Much more likely is someone buys their assets for 50 cents on the dollar or less and continues to bill you.
If they die from the AC being cycled from a half hour you have a problem, besides they would have been killed by the blackouts the system is meant to avert anyways.
Sure, they are cross-platform, if you run the EXACT java version on each platform that they were written against, move even one subversion off that and you often run into issues. This is a real problem when the supported version has known security holes but the vendor hasn't yet updated the app to support a newer version with the patch. Also it's a royal pain in the keister to keep up with the matrix of what version is needed by each app, I have eight different JRE's installed in my Citrix image to support all the different LOB and hosted apps we use.
Dude, most of the java code I've run into isn't even future proof to the next subversion of java. It's seriously a significantly larger headache for us than going from Office 2000->2003->2010 or XP->Win7.
Slashdot Mirror
The average rate of return for the NYSE over the last ~150 years has been 8% annually compounded, there is no other investment vehicle which compares. If you aren't playing the game you are losing. Heck after the crash in September 2008 I increased my 401k contribution by 50% because I will probably never see a return greater than the climb out of this recession. If I'm wrong it won't really matter because modern society will cease to function and so the wealth wouldn't do me any good in any investment except perhaps bullets =)
I don't know of any single day uptick approaching 10%, but there have been quite a few days in the markets history that large.
At this point that's the least of their worries, there are criminal charges waiting for someone including possible federal charges.
Yeah, the GP was talking about big time storage, to me that means SAN =) And the ram cache I was thinking of is battery backed for writes which system ram obviously can't do, and even for SSD's a bit of ram cache allows you to reorder writes which can greatly improve throughput and life.
With hard disk failure rates approaching or surpassing 50% within the first 6 months for some manufacturers and lots (regularly) I'd argue that we've reached the point where SSD system drives Make Sense.
Uh, quit buying cheap crap... My failure rate across 170 servers and two SAN arrays is less than 1.5% per year.
Actually the near to midterm future is tiered storage with fast ram cache first, then SSD's and finally big slow, cheap disks (or perhaps big fast moderate cost disk if you have a database too large to fit into SSD cache and your application can't stand the high latency of SATA). This is the design of the new Sun storage servers and also the design that Netapp uses (I'm sure there's others but those are the two big ones I'm aware of) and I think it's the way that makes the most sense since it most efficiently utilizes the expensive (on a $/GB) SSD resources.
IBM better on service that HP? Either your from a different world or you work for a Fortune10.
Ha, downtime to swap out hardware?!? That's the biggest reason to go VMWare, swap out servers or storage with zero downtime (vmotion and svmotion for the win). Heck with Windows Enterprise I can hot add ram without downtime (CPU's too but most stuff won't take advantage without a reboot).
.5 to 3MW/hr? That's a pretty big facility, the entire 500 person HQ facility for my S&P 500 company including our datacenter only has 650kW/hr of generating capacity.
Doh! unit conversion error, big mistake and yeah it makes a heck of a difference.
Dude, the Valdez was 11 MILLION gallons, at 50k barrels a day it will take ~ two months for this to be as big. I'd hope we can stop this is less than a month, let alone two. Now this will be bigger in that it will affect a larger area, and a more commercially significant area due to the amount of seafood produced in the region, but to make this sound like it's the end of the world is a bit overdone.
Nope, you haven't needed a physical exemplar of your invention in a LONG time (like 100 years).
Like Proliant, MSA, EVA, etc? HP turned the Compaq products into some of the best selling products in the market and became the #1 computer manufacturer.
Isn't it ten more years of September, aka the September that never ended?
Archie only predated the web by about 24 months and sucked by comparison, there's a reason it died off quickly.
Gopher was already waning pretty heavily by 97, its heyday was probably 93-95 because by late 1995 you had Windows 95 and Navigator making the GUI web very approachable.
Why not lower the cost to upgrade to the first version that supported a modern distribution method? Might bring in more upgrade revenue and lower your and your customers operating costs.
Virtual media FTW, on 99% of my server's that's provided by ilo, on the rest it's from the ipkvm.
Your right, Oracle, ADP, and a few other large vendors are incompetent, doesn't change the fact that in practice Java code is no more portable than well written c/c++ code.
Flash+Flashblock should maintain your sanity while not requiring you to load up IE, and with the new plugin container in 3.6.4 it doesn't even take down the browser if/when it crashes =) Oh and Flash 10.1 also significantly reduces CPU load by using GPU offloading if available.
Since upwards of half the cost of the system is installation I don't think a bankrupt SolarCity or their creditors is going to go around pulling panels off your roof. Much more likely is someone buys their assets for 50 cents on the dollar or less and continues to bill you.
If they die from the AC being cycled from a half hour you have a problem, besides they would have been killed by the blackouts the system is meant to avert anyways.
Sure, they are cross-platform, if you run the EXACT java version on each platform that they were written against, move even one subversion off that and you often run into issues. This is a real problem when the supported version has known security holes but the vendor hasn't yet updated the app to support a newer version with the patch. Also it's a royal pain in the keister to keep up with the matrix of what version is needed by each app, I have eight different JRE's installed in my Citrix image to support all the different LOB and hosted apps we use.
Actually, electric hot water heaters in the desert make zero sense to begin with, freaking install a flat panel collector already.
Dude, most of the java code I've run into isn't even future proof to the next subversion of java. It's seriously a significantly larger headache for us than going from Office 2000->2003->2010 or XP->Win7.