Society – Thin Client Model
Dec 14th, 2008 by sherri
A few months ago I walked into Radio Shack, looking for a short-range FM transmitter. I asked the woman behind the counter if the store sold FM transmitters.
“I don’t know,” she frowned. “The Internet’s down. I can’t access our product catalog.” (Gah!)
Weeks later, I walked into a U-Haul to rent a truck. The computers weren’t working properly, and the manager was having trouble completing my transaction. “What happens if the computers are down?” I asked. “Can you still rent me a truck?”
“Well, I can,” he said, “But that’s because I’ve been here for fifteen years and I remember how to use the forms. That kid over there–” he gestured toward the younger employee, “He doesn’t even know the paper forms exist.”
As communication technology advances, society has shifted from a thick client to a thin client model. Until recently, Radio Shack employees maintained product knowledge in their heads and on paper that they could physically access. U-Haul staff used paper and ink to rent out their trucks. Individual stores could operate independently of the central system, at least until supplies ran out. They each had to maintain up-to-date books and forms, and train employees.
More and more, information resides on remote systems, which distributed franchises and employees access in order to conduct transactions. On the one hand, this increases efficiency. Gone are the reams of preprinted contracts and forms to be manually filled out for each transaction. Employees have less to memorize, as information and procedures are built into software systems.
On the other hand, individual locations are increasingly vulnerable to network disruptions. Many businesses today rely upon the Internet in order access central databases and conduct normal transactions. Without connection, they’re just appendages cut off from the central body. Radio Shack may have FM transmitters, and U-Haul may have trucks, but without network access they have difficulty conducting business. Many businesses do not physically have the paper and supplies to support manual transactions, let alone the knowledge of manual procedures.
Do the benefits of the thin client model outweigh the costs? That depends on your perspective. From Radio Shack’s point of view, the vast savings from cutting employee training and paper supplies probably does outweigh occasional losses due to network outages. This is especially true if they create a more stable infrastructure than their competitors. Furthermore, in the thin client model, employees require less specialized knowledge, and are therefore more mobile (and expendible).
However, as a society our economic dependance on the Internet may be premature. The Internet was not designed for security, and as noisy worms have demonstrated, it can be brought to a standstill by small groups of people or even by accident. If a widespread network outage brought businesses to a halt, Radio Shack might not lose market share compared to other businesses, but society and the individuals within it would suffer.
The vulnerability of the thin client model was strikingly illustrated back in 2002, when Beth Israel Deaconess hospital “experienced one of the worst health-care IT disasters ever. Over four days, [the] network crashed repeatedly, forcing the hospital to revert to the paper patient-records system that it had abandoned years ago. Lab reports that doctors normally had in hand within 45 minutes took as long as five hours to process.” The emergency department was forced to close down and divert patients elsewhere.1
The disaster also helped hospital staff understand the benefits of the thin client system. One physician commented, “When I do this on computer, it checks for allergy complications and makes sure I prescribe the correct dosage and refill period. It prints out educational materials for the patient. I remember being scared. Forcing myself to write slowly and legibly…Without that dashboard of information I’d get from the computer, I had to walk up to patients I had treated before and ask basic questions like, What allergies do you have? Even if I thought I remembered, I didn’t trust my memory.”2
Will individuals become “dumb terminals”? Or will we simply evolve different kinds of processing capabilities? During the past few decades in the computer market, we’ve oscillated from thin clients to thick clients and back again. In the early days of computing, people used dumb terminals to access a mainframe, which stored and processed the data. Later, personal computers emerged, and each individual machine ran specialized applications and hardware.3 Nowadays, with the emergence of web-based business applications such as Google Apps and other client-server business processing systems, data is increasingly stored and processed on central systems once again.
Business processes will always mirror the technologies upon which they depend. As computers and business become increasingly intertwined, the efficiencies and vulnerabilities of our economy reflect those of our information technology. Humans have limited information storage capabilities, and leveraging centralized data storage systems helps us function as a group more efficiently.
How can we leverage the efficiencies of the thin client model, while still maintaining a robust and reliable infrastructure?
1Berinato, Scott, “All Systems Down,” April 2003, http://www.cio.com.au/index.php/id;1681249874
2Berinato, Scott, “All Systems Down,” CIO, April 2003, http://www.cio.com.au/index.php/id;1681249874
3Greenberg, Steve, “What Is Thin Client Computing?,” For the Record, July 2000, http://www.thinclient.net/technology/history-short.htm
Sherri Davidoff
Businesses are scared of volatility in all its forms, including variance in employees’ judgment and skills. Quality assurance requires (at least according to ISO) that processes can be repeated as accurately as possible independent of the people executing them.
When making systems less vulnerable to human error we often presume that computing systems become more reliable than humans can ever be. Large-scale IT outages caused by increasing system complexity will hopefully shift the focus back to people’s education as the major asset for companies.
This is something I’ve given thought to more than once. I’ve come to assume that someday, whether due to massive attacks on the system, solar flares, mismanagement, world war, or whatever form of disaster, a prolonged loss of the information superhighway to at least a good-sized chunk of the developed world would be a sufficient wake-up call to spawn investment into a completely separate privately funded Internetco(tm). Local service providers and larger businesses would tap into both nets, using one or the other as backup.
The costs of going back to paper-based systems simply to improve continuity of transactions would simply be too daunting and ‘backwards’ for large organizations.
Sherri,
I got 2 months of AC storage free from Uhaul because of a computer outage, and a nice manager. When we moved from Boston to Houston, I had called to arrange storage at a Uhaul location near the area we’d be staying (and moved to). When I went in to sign up, the manager told me that if I bought one of the locks from them I would get a month free. I kinda felt like it was a scam, but considering we’d just moved cross country any savings we could get would be great, and we figured we’d be out of the storage within the month anyways. We found a house and signed just after the first month, but they didn’t bill me because of a computer outage! I moved all of our items out of storage to our new place before the second month was up, and when I went in to pay up the manager just told us that we were set. Since we’d just bought a house and nothing was missing from our storage I was fine with this.
Anyways, Houston is not nearly as cool as Boston, but no snow (the snow we had last week doesn’t count)!
-miah
Sherri,
Great post! This is also a concern with physical security systems as more and more business place their physical security systems onto the network. I know of two instance where network outages took physical access control systems off-line, in both cases the outage was caused by a network worm basically DoSing the network.
In one case, at a large university, when the network went down the readers went into a fail safe mode and would not unlock the door. In this case they needed to use old fashion brass keys to open the door. The network outage occurred after hours so didn’t cause a big impact and only took a few hours to clean up.
In the second case, a large hospital, when the network went down the door all unlocked themselves. In this case the hospital needed to bring in a large number of guards to monitor door to sensitive area. All in all it took them a few days to get the network back online. I never got an estimate on how much the network outage cost the hospital but the guard cost alone were probably substantial.
So when moving to a thin client model it’s also important to test and understand what happens when the network go offline. Does the system fail in a graceful manner?
Cheers,
Matt
Ha! Check out the 2007(?) “VCF East 4.0″ interview with Chuck Peddle, famous of MOS Technologies / Commodore (and graciously uploaded to YouTube by Dave Haynie of the Commodore-Amiga team): http://www.youtube.com/user/hazydave
Somewhere in there is the anecdote of his participation in a major retailer’s (Macy’s?) decision to go electronic at some point in the early prehistory of computing. Of course, the day after Thanksgiving, someone hit a pole, and all the dumb terminals at the stores went out — reinforcing his belief that there’d be a market for a smart terminal… really a computer… a simple, personal computer… that’d have enough smarts to handle transactions locally and reconcile them with headquarters when possible — and not leaving an entire company dead in the water when “The Internet’s down.”
Hmm, one way to somewhat mitigate thin client problems is for each machine or at least each store to have a local cache of important info. The radio shack might know if they had any FM transmitters, but not if the store in the next town did. Individual med stations might have information for any patient currently in the wing, but not info on past patients or patients in other parts of the hospital. This makes theft of hard drives a more dangerous event, but keeps data in a more distributed, error resilient network.
Kevin H:
To mitigate the theft risk, why not encrypt the local storage devices and keep the decryption key on something like a flash drive locked in a manager’s safe?
The concept of caching data on machines like that is a lot trickier than you can imagine. If any of you have time, do some looking into “occasionally connected computing”. It’s been a major deal for “field agents” such as salespersons and technicians. Data synchronization between two “authorative” sources is highly complex. Beyond the data itself, the software architecture for accessing that data in a useful manner is also troublesome to support in a distributed environment.
Actually, some hospitals (like UPMC) are developing things of that nature. There are several pilot programs in place where PDAs are kept at the nurses station and patients records are both created and transferred to the PDAs. So as patients are evaluated, that evaluation is placed on the PDA and then ‘synced’ at the nurses station. In this scenario, you would have a series of devices that have all current patients information on them (but not past patients or patients in other wings).
However, even with keeping current patients information available in case of a network crash, you have an issue of patients checking in during the crash. As referenced in the one Dr quote above, he had to ask patients he had treated in the past about their medical issues because he didn’t trust his own memory. And that is probably the real danger.