Is the Physical Thin Client Era Dead?



  • This is because a thin client isn't really a thing, it's just a configuration.



  • @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    @StorageNinja said in Is the Physical Thin Client Era Dead?:

    @bbigford said in Is the Physical Thin Client Era Dead?:

    They are slow as fuck in most environments

    Are they slow, or did someone underprovision the Shitrix environment behind it?

    the problem I've always had with thin clients was flash. Any app or webpage that used flash caused the whole screen to flash white between pages. Though this never happened on a typical PC - Even Windows XP with 1 GB RAM - it never flashed and worked well.

    You mean you had a problem with RDP or other remote sessions, not Flash locally on a thin client? I think you are mixing the concept of the hardware with the effects of some remote access protocols. Very different things. It's like being unhappy with your car based on not having found a road that wasn't congested.

    It was RDP in both cases into TS/RDS. The only difference was the hardware. So yeah, pretty sure I was comparing only the cars on the same road.

    Had to be different software handling the protocol. You used the same RDP connection, one from a "full PC" and one from a thin client to the same RDS server and got different results?

    This suggests that your thin client was likely not up to date (many are not, a standard problem with them) and that it was falling back to an older RDP version.

    While is suppose it’s possible, the last testing I did was with a brand new HP thin client latest software versus an XP PC. Same problem, flag based apps/pages would always flash white on the screen.



  • @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    This is because a thin client isn't really a thing, it's just a configuration.

    My thinking of the time was that the RDP protocol was able to use the video card abilities better in the PC versus the thin client.... ie the thin client had shit hardware for the video card/component.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    @StorageNinja said in Is the Physical Thin Client Era Dead?:

    @bbigford said in Is the Physical Thin Client Era Dead?:

    They are slow as fuck in most environments

    Are they slow, or did someone underprovision the Shitrix environment behind it?

    the problem I've always had with thin clients was flash. Any app or webpage that used flash caused the whole screen to flash white between pages. Though this never happened on a typical PC - Even Windows XP with 1 GB RAM - it never flashed and worked well.

    You mean you had a problem with RDP or other remote sessions, not Flash locally on a thin client? I think you are mixing the concept of the hardware with the effects of some remote access protocols. Very different things. It's like being unhappy with your car based on not having found a road that wasn't congested.

    It was RDP in both cases into TS/RDS. The only difference was the hardware. So yeah, pretty sure I was comparing only the cars on the same road.

    Had to be different software handling the protocol. You used the same RDP connection, one from a "full PC" and one from a thin client to the same RDS server and got different results?

    This suggests that your thin client was likely not up to date (many are not, a standard problem with them) and that it was falling back to an older RDP version.

    While is suppose it’s possible, the last testing I did was with a brand new HP thin client latest software versus an XP PC. Same problem, flag based apps/pages would always flash white on the screen.

    Many thin clients don't run Windows and use a completely different RDP library. That is often the cause of issues.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    This is because a thin client isn't really a thing, it's just a configuration.

    My thinking of the time was that the RDP protocol was able to use the video card abilities better in the PC versus the thin client.... ie the thin client had shit hardware for the video card/component.

    Even in 2013 HP was shipping quad core i5 procs and serious graphics cards in thin clients.

    https://www.pcmag.com/article2/0,2817,2423364,00.asp


  • Vendor

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    Many thin clients don't run Windows and use a completely different RDP library. That is often the cause of issues.

    Many thin clients don't use RDP. While the protocol wars were fun (ICA vs. PCoIP!) I'm seeing everyone consolidate on highly customized stacks that at their core for image processing leverage H.264, and H.265. This is because for mobile and SOC have hardware decoders for this. (Blast Extreme and Citrix's HDX whatever it's called now do this).



  • @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    This is because a thin client isn't really a thing, it's just a configuration.

    My thinking of the time was that the RDP protocol was able to use the video card abilities better in the PC versus the thin client.... ie the thin client had shit hardware for the video card/component.

    Even in 2013 HP was shipping quad core i5 procs and serious graphics cards in thin clients.

    https://www.pcmag.com/article2/0,2817,2423364,00.asp

    Sure it might have been an option, but as previously mentioned - who wants to spend as much on a thin client as you do on a typical desktop? I suppose for a test I could have tried that, but alas we never did - we went with baseline units with onboard graphics... but then again, the XP IBM PCs we were running from 1999 were also all onboard graphics - yet never had the 'flashing' problem the thin clients did.

    I even worked with HP support in an effort to resolve the problem, which we never did.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    Sure it might have been an option, but as previously mentioned - who wants to spend as much on a thin client as you do on a typical desktop?

    Well, you do because of needs you defined.

    The problem is, you are comparing "fast hardware to slow hardware" but reporting it as "fat clients and thin clients."

    You are getting your correlation wrong. It has nothing to do with them being thin clients. It has to do with a trend of your specific experience in having purchased (or tested) thin clients incorrectly but fat clients correctly. It is your emotional reaction to how you perceive pricing "should be" for one and not the other, rather than how it actually is.

    If you want to do things that require heavy processing on the thin client, then you need a more expensive thin client. Period. It's that simple.

    And in general, thin clients are more expensive, not less (purely a market thing), so people wanting them "cheaper" is a misconception entirely.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    I suppose for a test I could have tried that, but alas we never did - we went with baseline units with onboard graphics... but then again, the XP IBM PCs we were running from 1999 were also all onboard graphics - yet never had the 'flashing' problem the thin clients did.

    I even worked with HP support in an effort to resolve the problem, which we never did.

    My guess here is that you were testing Windows fat clients against non-Windows thin clients with totally different RDP code. And likely that was the issue more than anything.

    The problem is here, you are testing different hardware, software, and approaches, all at once. Unless you do more testing to narrow it down, you can't really say that you had a thin client problem (because you proved that that can't be the case with part of your test) because you were testing two other, far bigger, things at the same time and there is every reason (and effective proof) that they, not the thin client structure, was the issue.



  • @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    I suppose for a test I could have tried that, but alas we never did - we went with baseline units with onboard graphics... but then again, the XP IBM PCs we were running from 1999 were also all onboard graphics - yet never had the 'flashing' problem the thin clients did.

    I even worked with HP support in an effort to resolve the problem, which we never did.

    My guess here is that you were testing Windows fat clients against non-Windows thin clients with totally different RDP code. And likely that was the issue more than anything.

    The problem is here, you are testing different hardware, software, and approaches, all at once. Unless you do more testing to narrow it down, you can't really say that you had a thin client problem (because you proved that that can't be the case with part of your test) because you were testing two other, far bigger, things at the same time and there is every reason (and effective proof) that they, not the thin client structure, was the issue.

    Point taken - I never believed the 'thin client' itself couldn't do it - Just wasn't expecting to need thin client devices as powerful or more so than desktop machines running XP to be required to get an as good experience. This whole thing once again showing that that setup is rarely about cost savings - which was the main reason we were even looking at thin clients in the first place.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    @scottalanmiller said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    I suppose for a test I could have tried that, but alas we never did - we went with baseline units with onboard graphics... but then again, the XP IBM PCs we were running from 1999 were also all onboard graphics - yet never had the 'flashing' problem the thin clients did.

    I even worked with HP support in an effort to resolve the problem, which we never did.

    My guess here is that you were testing Windows fat clients against non-Windows thin clients with totally different RDP code. And likely that was the issue more than anything.

    The problem is here, you are testing different hardware, software, and approaches, all at once. Unless you do more testing to narrow it down, you can't really say that you had a thin client problem (because you proved that that can't be the case with part of your test) because you were testing two other, far bigger, things at the same time and there is every reason (and effective proof) that they, not the thin client structure, was the issue.

    Point taken - I never believed the 'thin client' itself couldn't do it - Just wasn't expecting to need thin client devices as powerful or more so than desktop machines running XP to be required to get an as good experience. This whole thing once again showing that that setup is rarely about cost savings - which was the main reason we were even looking at thin clients in the first place.

    It can be about cost savings, but not at the device level. It's about lowering the cost of support and maintenance.


  • Vendor

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    Just wasn't expecting to need thin client devices as powerful or more so than desktop machines running XP to be required to get an as good experience

    Zero Clients have less intelligence than a rock (It's an ASIC that gets it's firmware by PXE boot) and I can play Skyrim on them over PCoIP. Printer Redirection will not really work, and god help you with a IO USB device over the WAN but a Thin client doesn't need to be that powerful for graphics beyond 2D resolution support, and number of monitor support.

    All the new Thin Clients protocols are based on H.265. That is decoded in cheap(ish) SOC. Even an old iPhone 5/iPad 4 support H.265. In this case the CPU load is zero for the graphics as it's fully offloaded end to end.



  • @StorageNinja said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    Just wasn't expecting to need thin client devices as powerful or more so than desktop machines running XP to be required to get an as good experience

    Zero Clients have less intelligence than a rock (It's an ASIC that gets it's firmware by PXE boot) and I can play Skyrim on them over PCoIP. Printer Redirection will not really work, and god help you with a IO USB device over the WAN but a Thin client doesn't need to be that powerful for graphics beyond 2D resolution support, and number of monitor support.

    All the new Thin Clients protocols are based on H.265. That is decoded in cheap(ish) SOC. Even an old iPhone 5/iPad 4 support H.265. In this case the CPU load is zero for the graphics as it's fully offloaded end to end.

    As mentioned by others - I'm sure it was the implementation of RDP in the devices I was using - from my POV it was passing the Flash video information to the end device for execution instead of a video stream. I can't explain it beyond that.

    With Flash being all but dead today, with any luck that problem is gone.



  • @Dashrender said in Is the Physical Thin Client Era Dead?:

    @StorageNinja said in Is the Physical Thin Client Era Dead?:

    @Dashrender said in Is the Physical Thin Client Era Dead?:

    Just wasn't expecting to need thin client devices as powerful or more so than desktop machines running XP to be required to get an as good experience

    Zero Clients have less intelligence than a rock (It's an ASIC that gets it's firmware by PXE boot) and I can play Skyrim on them over PCoIP. Printer Redirection will not really work, and god help you with a IO USB device over the WAN but a Thin client doesn't need to be that powerful for graphics beyond 2D resolution support, and number of monitor support.

    All the new Thin Clients protocols are based on H.265. That is decoded in cheap(ish) SOC. Even an old iPhone 5/iPad 4 support H.265. In this case the CPU load is zero for the graphics as it's fully offloaded end to end.

    As mentioned by others - I'm sure it was the implementation of RDP in the devices I was using - from my POV it was passing the Flash video information to the end device for execution instead of a video stream. I can't explain it beyond that.

    With Flash being all but dead today, with any luck that problem is gone.

    I imagine it was just really heavy video processing being sent without the benefits of "Windows compression" due to a lack of Flash awareness or something like h.264 that would normally handle it gracefully.



  • @StorageNinja said in Is the Physical Thin Client Era Dead?:

    @bbigford said in Is the Physical Thin Client Era Dead?:

    They are slow as fuck in most environments

    Are they slow, or did someone underprovision the Shitrix environment behind it?

    Not a configuration issue with infrastructure (Citrix or VMware), since zero clients ran great. The thin hardware endpoints were always just slow as fuck.