Device Management will become increasingly important as commercial enterprises, government agencies and public organizations roll-out Internet of Things oriented services that change business models, work-flows and supply-chains. The problems associated with effectively and adequately managing gateways, routers, end-devices and sensors, increase dramatically as the number of deployed devices increases and as geographically diverse locations are involved. The complexity rapidly reaches levels where manual or analog control of devices are no longer viable. Only robust management systems can provide effective use and maintenance of network assets. As a general rule. approximately 5% of the total cost of the network equipment deployed needs to be factored into the operational cost of such a network on a annual basis. Without which, the cost of network disruption or additional equipment will be 10 to 20% of the same factor. The choice is clear: Plan for a 5% operational budget for network management, or put 10 to 20% of the cost of your deployed equipment at risk every year. What should this 5% budget be applied to? My suggestion is to use the following as a model for investment.
The Problem: Device Management Complexity
For the purposes of managing things, complexity can be summarized as:
- the number of devices,
- multiplied by the number of functions or data-points per device,
- multiplied by the number of network connections,
- divided by operational utility (OU)
Of the four different variables, OU is the most difficult to determine. OU is an attempt to account for the value of the effort applied by IT or network personnel for management of network devices without a management system. Although an imperfect factor, we will establish that OU equals 1,000 devices per one full-time network operations technician divided by the number of locations (e.g. physical addresses) the devices are located at.
By using this algorithm a result of greater than 1 is an indication that the complexity of adequately managing the network has exceeded the capacity of the IT staff. The baseline example would be:
- Number of devices: 1,000
- Average number of data points per device: 1
- Average number of network connections per device: 1
- Number of full-time network administrators: 1
- Number of devices locations: 1
In the example the complexity factor is calculated as: (1,000 x 1 x 1) / (1/1) x 1,000 = 1 This validates the equation based upon the assumption that one full time IT or network technician can adequately handle management of 1,000 devices in one location. Other use-case examples produce the following results:
- Scenario #1: by increasing the average number of data-points to a more realistic 5, pushes the complexity ratio to 5.00; thus indicating the complexity of the network has reached a level beyond the capacity of one full-time network administrator.
- Scenario #2: the staff of network administrators is raised to 10 and the complexity ratio drops to 0.5; indicating there is over-coverage of the network administration staff.
- Scenario #3: illustrates the optimal complexity factor.
- Scenario #4: illustrates how adding additional locations increases the complexity ratio.
- Scenario #5: is based upon a real life example of an energy utility that has deployed an Energy Management Solution in 20,000 customer homes. Even with a dedicated staff of 100, the complexity ratio skyrockets to 200,000 times the optimal ratio. In fact, to meet the optimal ratio, the utility would have to employ a staff of 20,000,000! This obviously will never occur in a real world business model, but this does illustrate how quickly the complexity factor can elevate as the variables rise.
The higher the complexity ratio, the greater risk of network errors that cost IT, and ultimately the company, significant sums of money due to lost business costs associated with correcting the errors. As Internet of Things solutions emerge and grow, simply adding staff to achieve complexity factor optimization quickly becomes untenable. The only way to change the equation, literally, is to integrate management capability into the solutions. A management solution typically includes the core capabilities to remotely access and control devices. However, even these powerful capabilities cannot address the levels of complexities of large-scale networks.
To address the growing need for management, solution providers can take advantage of the capabilities best characterized in the “The Hierarchy of Need, for Management of Large-Scale Networks.” More on this in my next post.