- Jon Gaasedelen
This is the second in a series of tips exploring cloud computing in healthcare. The first tip in the series looked at cloud services in general and the uses of network-attached storage. In this tip, the uses of storage area networks are explored. Read the first tip here.
Storage area network storage, or SAN-attached storage, is a critical kind of storage to an organization. Its main components are storage arrays approximately the size of a large commercial air conditioner. These arrays are connected to multiple connection hubs (called switches), which are connected to servers through multiple input/output ports.
Redundancy at all points is a best practice for a SAN because it reduces the risk of failure. This concept is very important because a SAN-attached storage server writes data directly to an external hard drive through a maze of wires and cables. Whenever a server writes data to disks, it has to know that the data was written successfully to the disk before it will proceed to other operations. This protection is a carryover from the earliest days of computers, when quality of storage media was the weakest link in a computer system.
Most industries have implemented and maintain private clouds for their own operations.
Quality isn't an issue these days, but because there is still concern about writing operations to disk drives; so much effort is made in SAN-attached storage to acknowledge every write operation. With millions of write operations to disks taking place every second, this isn't a task to be taken lightly, and it's usually the largest cost factor in storage arrays. A realistic cost estimate is upwards of $500,000 for a small SAN-attached storage array.
With the expense and risks associated with SAN storage, one might ask, "Why bother?" The answer is speed. Even with a maze of cables and wires connecting servers to disks, the data rate (the speed of data flow) approaches nearly the speed of light because of the fiber optic lasers that transmit the data.
Another important note about computer systems is that all data must be stored in a format for writing and reading. This includes all data for all computer operations. Consequently, disk access (reading and writing to and from disk) can be a very limiting factor, so much so that computers' memory and CPU are implemented to make disk access much faster. Nonetheless, disk access is still a major cause of system slowdowns.
Disk access can be slow for many reasons, but to oversimplify: Access is usually the time taken to acknowledge a disk read or write. The system won't return control of the computer to the user until a read or write is acknowledged. Thus, computer response time and disk access time are important to patient care, particularly when the user is a physician triaging patients in the emergency room.
One thing not yet mentioned is how system storage can be accomplished using cloud services. A good visualization to use when you're picturing cloud storage services is to think of the cloud as existing in a remote, self-contained brick-and-mortar structure. This structure contains a server and a storage array. The array and server are both connected to multiple switches and data is flowing between the server and storage. Back at the hospital, nurses and doctors are logged onto the hospital network either through a desktop attached to a LAN or through a Wi-Fi capable phone or tablet.
Whether by a phone, desktop or tablet, users access an EHR application installed by IT and located on a server, usually in the hospital basement. Using configuration mechanisms and other IT practices, the local hospital application accesses disks located in a nearby brick-and-mortar building that holds the standalone server and storage array. This building can hold other hardware equipment that provides other cloud services, thus it can be considered to be a tactile representation of the cloud.
The concept of selling primary services (Infrastructure as a Service, Platform as a Service and Storage as a Service) is the only aspect cloud that's new. Most industries have implemented and maintain private clouds for their own operations. This means they provide all the computer servers and services for their own employees.
What is new is providing these services to the public in the public cloud for either excess-capacity needs or for services to small organizations that might not be able to afford their own private services. This is similar to a "lease or buy" decision. Using the public cloud for EHR software services or other services that meet communal needs is a realistic and economical option for many independent healthcare providers. A hybrid cloud option or a combination of public and private cloud also could be an economical choice for larger group practices of/or clinic and hospital partnerships.
Continue to the third part in this series.
About the author:
Jon Gaasedelen is an independent IT consultant with more than 20 years' experience in information systems infrastructure. He has an undergraduate degree in economics and a master's degree in health informatics, both from the University of Minnesota. Let us know what you think about the story; email firstname.lastname@example.org or contact @SearchHealthIT on Twitter.