Are thin clients a meme?

>>58354794

My uni uses them, "upgraded" last year. Absolute trash. Slow laggy fuckers with a 15 minutes login time.

I am pretty sure my uni it is pretty much the worst in the country.

>>58354794
Yes, they have shit performance i.e lag and they cost as much as a proper computer.

If they're used as intended so as a vnc client then they are pretty ok for basic office tasks.

>>58354794
yes
diskless setup>thin clients

The purpose of the technology is to improve the experience for the development/administration team, not to improve the experience for the end user.

Retail loves them because they can lock one server behind a door in the back office and serve the point of sale to the dozens of thin clients up front without worrying about exploits/etc.

But that's the only benefit of the tech. Physical security, and ease of administration. There's no benefit from the user experience or any other performance metric.

File: .jpg (152KB, 498x375px) Image search: [Google]

152KB, 498x375px

>>58354794
More like blast from the past. Literally pointless when these things exist. But storage on the server is a good idea.

>>58355009
You can make a thin client from this machine.

>>58355484
You can make a thin client out of anything. But why? This runs full desktop just fine.

>>58354818
>>58354841

That's a server problem, not the thin client's fault.

>>58355484
I won't.

>>58354818
ELTE?

>>58355728

Also almost all of the thin client setups I've seen are severely bottlenecked by underpowered server, underpowered network infrastructure, and often run by teams who have no idea how to work around those limitations.

I've seen shops who try to run full WIndows desktop environments on a dozen thin clients powered by a single server that didn't even have a GPU. I've seen a shop try to power two dozen thin clients off a server with a 100MB switch. I've seen shops try to roll their own thin client architecture and end up, very very poorly, reinventing existing FOSS tools .

>>58354818
you are either that tool who made up those dumb threads stating the '15 minute login time' to try and bait replies, or you are lying

>>58354794
they're good for POS stuff or menial work (warehouse inventory, etc)

for desk work though, just buy workstations

>>58354794
Right now since the Intel Atoms no longer performs like Pentium IIs, this technology is a meme.
>>58358416
That is for POSReady to handle

>>58354794
Had a displeasure using a Sun Ray "work" "station", definitely a meme. People in the past had it even worse with mainframes.

>>58358807
i have no experience with oldschool terminal and server setups.....well, except for when i was in middle school and my mom was working at an office where she used a straight up text only amber terminal in the mid-90's. it i don't seem to remember any lag when i watched her using it. how bad was lag and slowdown back then? the few times i've touched a thin client it was slow as fuuuucccccck.

File: middle ground sucks.gif (14KB, 349x336px) Image search: [Google]

14KB, 349x336px

>>58354794
Sadly it seems that they will stay for awhile longer, so I guess they are not a meme.

I do however wonder who remembers how a Mainframe is connected to a Switchboard which is connected to Terminals.
Because it seems oddly similar to how a Sever is connected to a Hub which is connected to Thinclients.

There is a reason people started buying desktops/workstations, they get shit done.
Sure sharing power and resources can be more efficient, but given the life cycle cost I would just pay the difference and get my own private workstation rather then bother with network maintenance and everything else.

>>58358903
It wasn't about lag. It was just extremely low performance single task machines, and you literally couldn't use them whenever you wanted to - they used time sharing.

>>58354794
Great when the infrastructure is good and you really need to lock down everything and conserve licenses. Often not cheaper than traditional fat client.

>>58354889
>diskless setup>thin clients
WTF do you think thin clients actually are?

>>58354794
No, not when done right.

The problem is most people don't do them right.

You can't over-subscribe the VDI infrastructure like you can your server infrastructure. I know it sounds weird, but a server can wait ~50ms for something where it just shows as connection lag in the user experience.

On a typical VDI host (32c/64t) we assign the VM's 2c / 8GB RAM / 80GB SSD storage, and build about 40 of them on the host. There are network shares for user profiles as well, but those aren't as quick because they don't need to be.

We also enable GPU acceleration, and I've found that the AMD FirePro V5800's are about the best bang for the buck at the moment, largely because they're slot powered and single slot width.

Just get NUCs for fuck's sake.

>>58359145
typically they boot an OS from an internal disk, but are only used to access remote applications, such as via VNC/RDP/Citrix

a "diskless setup" would typically mean a normal workstation, where applications are running locally, but all IO is remote, such as by booting via PXE, and using an NFS share as your root rather than mounting a local disk

of course, they're not exclusive, you can have a diskless thin client (no local disk nor locally run applications)

>>58354818
not needing any kind of performance is pretty much the point of thin clients

they can be shitty, cheap ass, old boxes, and it won't matter, because everything is happening on a remote server, the thin client is just a graphical terminal

if it's slow, then it's a network or server issue

so I have a question: if you connect a workstation to a terminal server with RDP and you access applications through the RDP, is that called a thin client environment? or a remote session?