Data silos are often used as the means for storing an organization’s data, having become one of the primary storage methods decades ago and surviving until today. A data silo (sometimes used interchangeably with information silo) is a repository that remains under the control of one department in an organization, remaining isolated from the rest of the organization. This is comparable to how a grain silo on a farm is closed off from external elements. However, data silos have undergone scrutiny in recent years, and some feel that they are a hindrance to the growth of the Internet of Things (IoT) and must be removed.
One of the key elements of the IoT is connectivity. Data from a smart phone or from the billions of sensors that have been placed in industrial workplaces is able to be tracked in real time because of its cloud-based connections. However, data silos notably retain data in a state inaccessible to outside sources.
As stated in ISO/IEC 17788:2014 – Information technology – Cloud computing – Overview and vocabulary, “cross cutting aspects are behaviors or capabilities which need to be coordinated across roles and implemented consistently in a cloud computing system.” These include availability and interoperability. As companies begin to build technologies that facilitate the flow of data by integrating with other platforms, they will face the large hurdle of the existing infrastructure that cannot meet this need.
Data silo shortcomings are not present only because of integration with the IoT, but have been brought up for other general reasons. Data silos have been criticized for impeding collaboration between different departments in an organization, even provoking competition among them. This, in turn, can lead to a loss of productivity and efficiency. Through a lack of information sharing, for example, a central database for a clothing retailer might include “extra large,” “XL” and “TG” as titles for the same size.
For analytics, this limits the data just to the perspective of a single department. This data comes from a very limited number of sources when it could very easily be a collective effort. While this ultimately influences the organization’s decisions, it is surely less reliable than integrated data.
Another complaint is the storage limitations that are faced by using a data silo. If data center requirements increase, the silo has to be scaled. This requires data center design, development of an integration plan, procurement, solution implementation, and, once complete, long-term management of the expanded silo.
One of the drastic cases of problems caused by data silos has been in the healthcare industry. Traditionally, hospitals maintain a significant amount of information on their patients, which they do not share with other locations, even if they are in their same area and could easily have the same patients. According to economists, the rationale behind this is that the patients are their customers, and sharing their information would involve giving their customer list and information to their competitors. However, data integration, or even just data sharing, between hospitals could enhance diagnoses and treatments, and the ease of access to data could save time and money.
Since data silos can impede the production of single organizations, in which each department has their own data silo and shares little information, it is very plausible that their architecture could limit the growth of the massive Internet of Things. The data accessibility of the IoT is practically the opposite of that from a data silo; it is constantly receiving data from a variety of sources to encourage the best decisions. Removing data silos might just help.
Cloud Computing Standards are available on the ANSI Webstore.