I can recall from my youth a day where my cousin and I hung out at my uncle’s engineering office on a Saturday. Maybe it was several days, but as I get older the days seem to meld together.
While my Uncle worked on whatever project was eating up his weekend, he placed us in nearby cubicles and logged us into the facilities mainframe computer.
I found myself lost in countless hours of Zork.
For those of you not fortunate enough to have experienced Zork, it was one of the earliest interactive fiction games available for computers.
They just don’t make games like that anymore – for better or worse.
Shifting back to present time, it dawned on me as I was presenting yet again on a thin and virtual architecture I put together for a client, that everything old was new again.
We are once again supplying operators terminals to access a cabinet full of computing resources. And for several good reasons.
In today’s data driven world, companies are looking for solutions that are fault tolerant while being easy to maintain and configure. At Gray Matter Systems, we have developed what we like to call the Virtual and Thin Architecture.
By being Virtual and Thin, what we are doing is leveraging Thin Client Technology in a VMWare Environment. This gives us several advantages.
What is Virtualization?
When I said the word server, most people would think of a physical box, by virtualizing that box, we can host many software based servers within it. This reduces our overall footprint on several levels – physical space, power consumption, and unused resources.
Using VMWare also allows us to create a Highly Available and/or Fault Tolerant environment by leveraging VMWare technologies such as VMotion, VMotion High Availability, and Fault Tolerant.
VMotion allows you to manually move servers from running on one host to another dynamically while the server remains running.
VMotion High Availability automates this process such that when a Host Box fails, any server that was running on said Host Box will restart on another Host Box in the Virtual Center.
Typical reboot time for a Virtual Server is around 30 seconds as compared to the 5-10 minutes of a typical physical server – thus minimizing downtime.
Couple that with iFix Enhanced Failover and the end user will not see a loss of data as the redundant server would take over operation as the primary server automatically restarted on another host box. Within 1 minute you’d be fully redundant (hardware and software) again. VMWare fault tolerant can take the hardware redundancy one step further and start a shadow instance of a server on another Host Box in the Virtual Center.
In this case, when a Host Box fails, any server that was running Fault Tolerant that was on that host box would see its corresponding shadow go active near instantly – thus no downtime at all.
With our servers centralized, we can now look at how our clients connect to the data contained in our databases (which are running on the servers).
For the last 20 years, thick clients have been the prevalent technology used to connect to our data servers. This required a physical PC to be located and maintained at each operator station or any station that needed access to our data. We now recommend the use of Terminal Servers and Thin Clients.
Thin Clients Are Not PCs
They are devices that generally have no moving parts that power a keyboard, mouse and monitor and establish an RDP session to a Terminal Server (Citrix or Web as well). The actual processing and configuration for the clients are then maintained on the Virtualized Terminal Server running in our VCenter with Virtualized Data Servers.
Thus all of our computing resources are centrally located and maintained. Replacing a Thin Client can be done in a matter of minutes.
Even less if you use Thin Client management software such as ACP’s Thinmanager.
Our typical rollout these days for a distributed client server SCADA system consists of the following:
- Redundant set of iFix SCADA servers
- Redundant set of Terminal Server running iFix Terminal Server and ACP Thin Manager
- Proficy Historian
- Optional Proficy Webspace
These servers in turn are then accessed by thin clients out on the plant floor as illustrated below:
The overall solution provides for ease of maintenance and a centralized configuration. All the while allowing you to more efficiently use your processing and storage resources while using less power and having a smaller overall footprint.
Lastly, it provides significant uptime performance and fault tolerance.
It’s funny how as times change, aspects and strategies of the past become vogue again.
In many ways we are back to that “mainframe serving the client terminals” architecture of the past. Even bell bottoms made a comeback in the near past. As far as myself, I’m just waiting for someone to develop an open-world version of Zork for the PS4 or XBox.
One can hope.