Last week, Microsoft introduced its Windows 365 a simplified version of its ‘virtual PC’, which is also availabe via its Azure Virtual Desktop architecture. For an interesting take on the concept, check out this article by Bob O’Donnell, an occasional contributor to Display Daily. It got me thinking about thin clients and virtual desktops. Again.
Terminals for mainframe and mini-computers were early consumers of high quality displays but these were, basically, character driven and lacking graphics. The move to GUIs with environments such as Gem, XWindows, Windows and Linux changed the challenge for terminal makers. Rather than just sending streams of Ascii back and forward, bandwidths needed increased. However, the vast increase in bandwidth has basically seen to that challenge.
I first came across reports of the Wyse ‘Windows Terminal’ around Comdex 1995, but I had a demonstration of the device at CeBit in 1996. The system allowed Windows Apps to be run on an Windows NT server, but remotely controlled from a simple graphics terminal. Communication between the WinTerm and the NT system used the Citrix protocol. NT was a good server operating system and the idea of being able to secure and maintain the desktop centrally always seemed like a good one to me. The Winterm used any monitor – Wyse made them in those days.
My company, Meko, developed a market research business in displays (since acquired by IDC) and at one time we had more than 20 staff collecting and processing data and producing reports. From the early days, we used Lotus Notes, which had really good remote working facilities, for many processes, including our newsletter publishing (and having access to the database back to 1994 is one of the reasons I can be sure of a lot of dates and developments). However, for data collection we used Microsoft Access initially, and later SQL Server, with a lot of processing and front end applications developed in house using VBA in Excel and Access.
Working from Home – Before the Pandemic
A lot of our workforce was part time and liked to be able to work from home, so we started to use virtual desktops hosted on servers at our office. That worked very well. Staff could use their own PCs for their own purposes and just needed to use the Windows Remote Desktop app to logon to the office. I was very proud that for many many years, I could do everything that I could do in the office from anywhere in the world. That needed access through the firewall router. What happened if that failed? I managed to find a power switch that was operated via the PSTN telephone network. I could dial in and re-boot the router if I couldn’t talk to the network. I even found a power strip that was internet configurable so that I could check on all the servers (we had a few by then).
(It only failed once. One winter, I was at CES in Las Vegas and there was a severe storm in the UK that caused a power failure. When the servers came on, they caused the circuit breaker for the office building to trip. I couldn’t fix that and my long suffering wife trudged the 5 kilometres from home to the office, to flick the switch again.)
One day, I realised that I was paying a lot of money for offices that were really only serving to act as a meeting place, and staff were having to spend time and money going to the office. So we closed the office and the servers moved to the garage in my house. That saved a lot of money and the business really didn’t suffer. I had a spare room in the house that could be used for meetings and for occasional jobs that really needed access to the servers. The world has caught up with this in the last year and a half.
Over the years, only one member of staff – based in the US and a Mac user – ever had persistent issues in connectivity. She did a lot of work with local IT people and her ISP, but never found out why she had frustrating drop outs. At one time, she was having trouble with connectivity, but at the same time I had a perfect connection from a remote office in Rio de Janeiro for a couple of days when I had to process the quarterly data in the same week as I had been asked to keynote the SID LatinDisplay conference.
For a long time, I was a big advocate of the idea of moving to remote desktops. A display company that I know, that I think was influenced by my enthusiasm, developed thin client products that were very good. However, the harsh reality was that few buyers cared much about display quality. The firm developed a lucrative market for a while supplying auditors that loved being able to go onto client’s sites and have full secure access to their systems from simple light terminals with high reliability and quality. However, eventually, the firm gave up and left the market to (mainly) big companies such as Dell and HP. Dell, of course, bought Wyse back in 2012. Companies such as Igel in Germany and NComputing are exceptions to the rule that only large companies can succeed in the market.
Over the years, I was surprised that the thin client paradigm was not more widely adopted. I think there are a couple of reasons. One was that systems used by corporates often have a wide set of variations, with different requirements for software, but also hardware. Systems often have very different applications. Some may need GPUs or other special hardware. The challenges of having so many different variations of PC “images” were very well summarised in a book I read – “The VDI Delusion“.
However, another factor, I think, is that moving away from PCs on the desktop is a high risk strategy. It only really works well, I think, where almost all the PCs in the organisation are virtualised. Otherwise, you still have to deal with hardware generations and support. If it goes well, a change to virtualisation could have big rewards, but if an introduction goes wrong, it would be career limiting. It’s my experience that those that rise to the top of IT management structures, and who would have the power to introduce wholesale VDI, are rarely those that chase high risk/high reward strategies.
I had planned to talk about risk and buyer behaviour, but I’ll leave that topic to another day! (BR)
(It may be that the huge penetration of cloud computing has significantly reduced risk and/or the perception of risk in the last couple of years. I’m no longer deep in the corporate IT world, so it may be that this view is getting out of date).
