Various companies are betting that the Internet of Things will bring a lot of value for companies that can harvest the data spewed from various networked devices. The fundamental assumption behind this is that there is business value to be gained that far outweighs the cost of gathering, storing, retrieving and analyzing this information to provide actionable insights. Since there is a lot of data processing costs associated in this paper, we propose some ideas on how to build a value framework into the Device control protocol itself that helps us identify the device and selectively pick different levels of information for each device. We propose to extend any device control protocol and provide an inbuilt ability to discriminate operations that are necessary for the data, in other words, which data should be collected but not stored, stored but not retrieved, retrieved but not used for analysis, analyzed but cannot be acted upon leaving the end user with the most optimal subset of information that can be acted upon based on the analysis.Evaluating the Costs associated with the Internet of Things

When people talk about Internet of Things and the opportunity it presents, it is vital for the people who evaluate the opportunity to consider the costs associated with exploiting such an opportunity. In most cases these costs would be related to Information Processing: – Storage, Retrieval and AnalysisDTc = DSc + DRc + DAc + (Licensing or IP costs) (Total Costs of Data) = Data Storage costs + Data Retrieval Costs + Data Analysis costs + IP costs if necessary for analysis.

Notion of Value

The notion of value is different for different businesses and when we ask ourselves what business value can be obtained from analyzing device generated data, we segregate the value into two buckets   DTv  = DBLv + DTLv (Total value of Data) = Total Bottom Line value + Total Top Line value Improving Bottom Line Device data can be collected and analyzed to identify failure and value characteristics

  • Profile the device into different classes which allows cost reduction by reducing storage, retrieval and analysis costs
  • Prevent failure

Improving Top Line A pre-requisite for the top-line opportunity based on information is that there is a profiled class of devices which can be targeted for a revenue opportunity. Some examples of revenue opportunities which can result in the improvement of top line are:

  • Replacement opportunity
  • Break fix Service Opportunity with certain SLAs
  • Value adding Monitoring Opportunity with certain SLAs

Underlying Rationale Discrimination and Grouping of entities or objects by their nature is an important aspect of human intelligence that we try to mimic in Machine Learning and Other Big Data activities. This ability to discriminate and group entities or objects allows us to perform operations on these objects mentally and communicate with others without the need for additional detail.To illustrate, all of us have the ability to model and abstract things to make judgements.  One of the fundamental assumptions within that modeling and abstraction, is the ability to group things together and consider them all to have the same behavior. This is also the fundamental assumption in arithmetic when we add numbers (1+1 = 2). The reasoning is the symbol 1 represents some object or entity and with the operation of addition we mean that as a result of this operation we end up with two entities or objects. The ability to discriminate and classify items into groups of similar items is fundamental for decision making and getting insights about the data.

No two customers are alike / No two devices are alike

We all know that no two customers are alike. Many companies take this fact into consideration by discriminating their customers such as Key Accounts, Major Accounts etc in their reporting. However, we all know that even among the Key Accounts the needs are not the same and even though they buy the same product, they may be providing different value propositions for each of these customers. Also the same customer might have multiple devices all of which may not be put to the same use and may have slightly different purposes, thereby lending credence to the fact that devices in fact can be discriminated not just among customers, but also within a single customer.