When faced with the task of building a “system,” the only way to go is to build the most appropriate solution for the situation. In some cases, that may be a fat architecture; in some cases, it may be a thin architecture; or it may be a little “chubby” client (a hybrid of both), but the main point is to build the most appropriate solution.
Fat clients will not automatically be replaced by thin clients. Either approach has its share of positive and negative attributes. For a fuller discussion, please see the article, What are the pros and cons of fat and thin architectures, and will thin replace fat in the future?
The trend across many businesses regardless of industry is a move towards thin client systems, primarily because thin client systems can support on-demand and other Internet-based applications with relatively little administrative or technical support. If you want to operate in a thin client environment, you’ll need to make sure that your network resources are extremely robust and you have some form of guaranteed uptime since thin clients can’t do a lot of work when the network is down.
Many of the advantages to taking a thin client approach revolve around cost savings. Workstations running the thin client do not need the vast system requirements that the application may require. Because of this, it is possible to outfit the workplace with low-cost computers that do not have the fastest, newest processors, lots of memory, and storage space. Only the computers that are really running the actual application need to be expensive, state-of-the-art machines. This cost, the cost of the computer needed to run a fat client, can be a negative factor. For fat client situations, the desktops do need to be state of the art, multi-processor, high RAM machines which, in large enterprise situations can have massive cost implications. This cannot be ignored when making a decision to go fat or thin.
There are also cost-savings in license fees. Not every user of the application needs to connect to it at the same time. So instead of paying a license fee to have the application installed on every computer (and sit idle), you pay a license fee for every simultaneous connection to the application. Keep in mind that not all software vendors offer this option, so you will need to investigate that this is a licensing model option when choosing software.
Additionally, there are savings to be had with respect to time, which also leads to cost savings and increased productivity. When a new version of an application is released, or if there is a maintenance upgrade, there is no need to install the fix, patch, update, or upgrade on every workstation. Only the computers running the application need to have the software installed. The thin clients on the workstations connect just as easily to the new version of the systems as they did with the old. In a large organization, this promotes a greatly reduced installation and deployment time that can save hundreds of hours. Downtime is also reduced since multiple thin clients can access one upgraded version and get back to work as soon as that upgrade is completed.
Thin clients can run just as easily on laptops, tablets, desktops, smartphones, and a host of other devices such as smartboards, all with virtually no dependence on the actual OS of the device. This means that key personnel can access the application while out of the office, from various locations (whether on the other side of the facility, or the other side of the world), which can be especially useful in the case of EHR systems or simply for emergencies where your staff needs to be connected.
Not everything about thin clients is perfect; there are some disadvantages which must be weighed in when deciding the direction. As mentioned above, thin clients do require a stable network connection, whether that is the local network or the Internet. If a router fails or if the connection is disturbed for any reason, work can often come to a grinding halt. Responsiveness is also sometimes an issue. Even the fastest connections are not faster than a local machine. Internet lag time and network transmission speed affect the thin client application. There is always some delay as information is transmitted over the network, and this delay gets greater as the distance to the servers increases (particularly for internet traffic).
If you are a globally distributed organization, and the servers (and thus the application) are located in a different country, then besides the lag time due to distance, you may be faced with local laws and regulations that apply to the location of the application but not to that of the client. You may end up in a situation where certain control of the application is dictated to you and not something you can control.
Also with thin clients, unless they are properly load balanced with redundant failovers, etc., they do tend to create a single point of failure which can be catastrophic for a business if it does not have proper contingencies in place.
It is not unreasonable to predict that thin client computing is the future for business, especially when the thin client related technology continues to advance at a pace that begins to remove, or water down the disadvantages. For now, as I mentioned before, every organization needs to weigh the advantages and disadvantages in terms of its needs before taking the step towards one or the other.