Why do you think that?
Originally Posted by alttab
Most of what a child learns is from home, from parents and technology (web, social networking etc) Soft skills they will already have its the static and outdated curriculum that stifles a childs creativity and this has been proven by a academic technologist in India. The teaching of the core subjects have not changed for nigh on a 100 years, students sat in front of a teacher spouting the same boring stuff sending them to sleep then asking them to memorise and regurgitate when exams come round.
Teachers (Don't want to be cynical) are just highly paid baby sitters, some of which are in the job for the holidays and the good remuneration (I wouldn't want to be one as they also take a lot of s**t from the students).
Trends don't change in education i see the same lessons taught as they were to me over 30 years ago.
Time for change me thinks and remember you heard it from me first hehe! :)
I can certainly see learning from home become far more important in the future, however I don't see teachers and schools disappearing anytime soon. There is an awful lot after all that can't be taught or assessed sat in front of computer and as you have said schools are essentially day care and parents won't want to give up work and the goverment won't want to take that hit to the economy.
it's hard to know where to put VDI clients. Do you implement as part of a rolling refresh programme targeting certain areas that you know need renewal. So, if ICT suite A is 8 years old and needed replacing last year do you replace 30 fat clients with 30 vdi clients....could be a possibility if you've got a couple of spare 96GB quad core xeon servers that can be the beginnings of a VDI farm. Would you even take the risk with VDI for ICT suite A if you thought there might be a need to do full frame rate video work ?
Originally Posted by DrCheese
OR how about a learning centre or admin staff machines when they need renewing, you know that no-one has ever done anything too demanding. So you could replace machines reaching their end-of-life with VDI, i think some kind of VDI and BYOD in small quantities is preferable to a rolling refresh programme that purely consists of replacing an old PC with a new PC.
yeah, we know that current schooling is stuck in a time warp, but a lot of how we go about day-to-day business in the uk doesn't seem to take advantage of what technology can offer. I see the miles of slow moving traffic on the other side of the A road going into the city centre first thing of a morning, and i question how many people *really* needed to be in their offices for that time that morning or attend that meeting they're trying to get to. Some of them need to be physically present somewhere, but we know that technology makes a lot of this stuff possible, Skype, redirecting phone lines. So much so that many more people could babysit their own kids instead of getting schools and after school clubs to do it!!!
Originally Posted by bossman
Moving forwards Downes and Seimman's concept of connectivism is probably the best framework from which to approach education if e-learning is to succeed.
Anyway how did we even get onto this topic?
Reading back it looks like bossman!
If you want to talk about solutions for education you eventually have to ask what the problems actually are.
Originally Posted by Dave_O
[ETA - oh AND the forum law of thread drift.]
Very interesting thread, VDI really does seem to split opinion (although granted a lot of the hype is coming from the vendors selling it!)
I've been watching it as a technology for about 3 years now and analysing when the point will arrive that it hits that sweet spot of price \ performance \ features that makes it the accepted choice in the same way server virtualisation is these days.
Problems I've seen thus far...
1) Until recently VDI = SAN = costly. I can't see the point of taking the cost of infrastructure you can buy in stages (desktops) then spending the same \ more to move it to a huge initial cost by putting it in the server room.
2) Number of disks... as with all of these virtualisation-esque solutions IOPS are king... doesn't give great VFM if you then have to stack 20 odd SAS 15k disks to get enough grunt to provision 30 desktops on demand. Yes there's kludges around "boot storms" but it's not ideal. Replacing (cheap) SATA local PC disks with (expensive) SAS 15ks seems a bit pointless... affordable SSDs for this will be when we see the big shift imo.
3) Video performance... getting better with HDX, PCoIP etc but still seemingly not 100%
4) Licensing... not so much of a problem if you have EES (and therefore Software Assurance)... then there's the cost of the VDI software itself
Possible benefits if VDI done right...
1) No "image creep"... every machine reprovisions every time so you never get issues with viruses, Windows Update etc
2) Easy updates... new version of Office? Install on the Gold image and it's done
3) Smaller desk footprint... I really like the look of the AXEL Zero Clients, certainly frees up some space
4) Remote access... I do like the idea of the exact desktop being able to be accessed from home as opposed to a "nearly the same" Terminal Server... saying that I think the "Windows on an iPad" idea rather gimmicky and not that much of a benefit in the real world
How can VDI get there?
1) Easily expandable, pay-as-you-grow model... VDI in a box with the grid architecture looks a much better choice... start with 2 servers and scale up... lower initial costs
2) Affordable server SSDs... one of those will take the place of 20 odd HDDs but the prices need to come down...
With regards to what @gshaw has outlined I'll go through the answers to the problems hes seen In my opinion.
1. - Needing a SAN - Not always the case but to do it properly a san in a must. can be done affordably using Open-E Makes San affordable from decent spec Hardware. I Installed on last week as a testbed here on a Dl380 G4 with 15K scsi drives and seems to be rock solid. i've specced a Dl180 G6 with SAS drives in for a client using 1Gb iSCSI. Cost is a lot less and you can get some good bang for your Bucks (with replication etc)
2. Disks - is done properly doesn't need huge amounts of storage so not loads of disks.
3. not quite 100% but getting there with graphics offloading to the client device
4. Licensing - this is always the major downpoint for people as the cost can be see as astronomical if not justified correctly.
1) not with VDI in a box now it's a grid that uses DAS... always their main selling point. Even with your DL180 example that's still what, 4-5k? Add that to two decent spec servers and you're not far off 10k just to get started. That's the cost of 30 PCs before you've even put the software and thin clients in... ouch
Originally Posted by glennda
2) the disks aren't for storage but for pure speed, I remember a presentation showing you need about 20-30 IOPS per VDI machine you deploy... a 15k SAS disk can only do 180 (if you're talking random reads \ writes) so to get 30 odd machines provisioning at once you suddenly need a fair few of them in there. At current prices "disks is cheap" doesn't apply anymore and that's where I see part of the problem from a cost perspective... not to mention the power used for them.
Of course 1 SSD is up in the thousands of IOPS so with those the problem goes away in an instant :D
Ah I wasn't talking about VDI in a box but VMware my bad - but in terms of a SAN an Open-E solution is very good value!
Originally Posted by gshaw
No worries, good to have views on both products and all the available options :)
Originally Posted by glennda
Btw does your DL180 scream like a banshee the whole time it's on? Was really disappointed at how poor the fan control is on ours... HP didn't seem too bothered about providing a firmware fix for it at the time (mid 2011). Seems to be related to the PCI-e riser card but they give the same "we know, we don't care" response to everyone :mad:
In terms of SAN - Nexenta based is very good value, and can do all the funky SSD caching (stick in a couple of STEC ZeusRAM and STEC Zeus IOPS), dedupe, compression, thin provisioning, clones, snapshots etc. Backend storage can be on 7.2k RPM disks, and let (small number of very fast) SSDs give you the performance required with a non Oracle price tag. I think it's reducing the cost of the backend storage required for VDI that will help to start bringing it more as an option for larger deployment.
Originally Posted by glennda
They've also just released something that would be targetted at the VDI in a box type environment - http://www.nexenta.com/corp/nexentavsa-for-view
I'm in Buckinghamshire
Does anyone have a full vdi setup that I can come and have a look at in action.
We just appioneted new Ict company for a new build. There whole project Is based on vdi. It's what they can make the most amount of money on!!!!! Again no one has spoken to staff, not even the head is ict
As some of the other posts I keep looking at vdi but yet to be fully convinced so a site visit would good if anyone can help.
There's @jamesfed's VID in a box demo day on 6th June. Possibly close enough?
Originally Posted by ozydave
Cheers for the mention - still have space here :)
Originally Posted by pcstru