This is the third part in a series of tips exploring cloud computing in healthcare, including cloud services and applications. Part one looked at cloud computing as a storage solution, and part two explored the types of storage available for healthcare data. Part three considers the use of public and private clouds in storage applications.
While cloud-based storage was generally discussed as appropriate for healthcare, specific knowledge required to implement a solution was not mentioned. This is what will be explored here.
The best place to start is to ask the questions: "Who, what, where and how?" Hospitals are a good use case for public cloud computing storage in healthcare because of the many different challenges they face regarding data storage needs.
The first question to ask is, "Who are the hospital's main storage vendors?" Initially, this question may seem a little manipulative or oversimplified. In fact, it's an important question because all applications require data and all data requires storage. The question of the hospital's main storage vendors ties into what the hospital's main applications are, and where this data is stored.
Hospital operations provide information on a hospital's electronic medical records system, its picture archiving and communication system (PACS), and its health information management system, and, to a degree, the clinic and practice management systems they support. This knowledge also leads to the question, "What types of data and data storage does the hospital have?" All hospitals will have two primary types of data storage: network-attached storage (NAS) data and storage area network (SAN) data. One challenge with knowing the storage hardware type is that NAS or SAN storage can be used for any hospital application.
To clarify, data storage can be described in three ways: by size, by speed and by width; size refers to individual file size. Applications store data in files and files have two types: executable types and data types. Executables run logical tasks such as, "if this, do that"; data type files store data about a patient.
An important point about executable type files versus data type files is that executable type files should be stored as physically close to a computer's CPU as possible for faster access. Generally, standard practice is to store application executables on NAS storage because this kind of storage provides the closest access (i.e., highest speed) to the computer server's CPU. Data type files can be stored farther away from the CPU, but the challenge to store them efficiently becomes greater as they become larger. This is why SAN storage is usually used for data type files.
The third way to describe data storage is by its width. Wide storage is storage that supports the executables and data files for many applications. Every application has executable type (AKA runtime) files and data type files. Runtime files should always be stored close to the computer's CPU. Because of this requirement, many hospital applications have their own dedicated computer server for a particular application. However, this situation is not advised for many circumstances since it complicates data sharing.
Since we know that every computer application requires storage, the next questions every hospital should ask is, "Where does an application's storage live? Do the runtime files live on a separate computer acting as a server that also holds the data type files? Or are the data files separate from the application?" The truth is that neither case is necessarily the norm. This is why the "where" question is probably the most important question to ask because the answer affects the whole cloud solution, not just the decision to use cloud computing storage in healthcare. However, since this article is about cloud storage, for now all examples will be made for public cloud storage, even though other cloud solutions may be more appropriate.
Where data is stored is usually based on two things: available hardware components and software applications. Software applications tend to dictate where data can be stored because of the previously mentioned CPU issue. The available hardware components dictate where data is stored because they are usually implemented together with an application or software solution. A major concern with this style of solution implementation is that it often leads to a situation called data siloing, or the installation of a computer solution and its corresponding data that is only accessible by a particular application, meaning the data can't be shared.
Often, individual data components of an application are the same across applications. For example, most applications will contain a patient's name and identifying number. This is likely the same across many applications, but siloed data doesn't permit sharing. Also note that storage must be increased since the same data is saved multiple times with the various records. Another situation dictating data storage location is local hospital-vendor managed applications. This situation is sometimes used for EMR or PACS applications, and by definition are public or hybrid cloud solutions.
Continue to the final part of this tip series.
About the author:
Jon Gaasedelen is an independent IT consultant with over 20 years' experience in information systems infrastructures. He has an undergraduate degree in Economics and a Masters in Health Informatics, both from the University of Minnesota. Let us know what you think about the story; email [email protected] or contact @SearchHealthIT on Twitter.