The IoT is in position to become one of the greatest application performance management challenges faced by IT. APMdigest asked experts across the industry for their recommendations on how to ensure performance for IoT applications. Part 4, the final installment of the list, covering communication and the network.

Start with Top Recommendations to Ensure Performance for the IoT - Part 1

Start with Top Recommendations to Ensure Performance for the IoT - Part 2

Start with Top Recommendations to Ensure Performance for the IoT - Part 3

19. THE NETWORK

The network makes the application possible, so having a reliable, robust and secure network is the best way to ensure performance of IoT applications. There are a variety of technologies that are being attempted for IoT networks but many will eventually fall short because they will be unable to evolve to deliver the performance that will ultimately be required. Cellular delivers fast speeds at a premium cost for bandwidth-intensive devices such as mobile phones, but cannot offer ubiquity of coverage and the optimization of power consumption – both critical attributes for IoT. Relying upon low-bandwidth high-latency tower-based technologies is dangerous, due to architectural capacity limitations. By using a mesh architecture, one can optimize for cost, coverage, scalability and reliability. A modern mesh architecture also offers increased reliability as use of the network increases because each new device broadens coverage and enables alternative pathways for communication. Finally, security is a key element of any high-performance application, and a layered, defense-in-depth approach, should be leveraged.
Don Reeves
CTO, Silver Spring Networks

20. PEAK TRAFFIC

IoT performance depends on both the availability and ability for the application infrastructure to support the number of devices and types of communications in the IoT solution. IoT is more dependent on harmonic communications where updates are sent at regular and fixed intervals. The application infrastructure must be able to deal with the peak communication patterns during these intervals in terms of volume of data and number of simultaneous connections. This is not unlike the localized surge in traffic experienced during major sporting events that occur on a regular basis (i.e. NFL matches).
Frank Yue
Director of Application Delivery Solutions, Radware

21. SEPARATE FROM BUSINESS NETWORK

A large IoT deployment can overwhelm and clutter a production network, simply in the number of devices and addresses used. Consider putting IoT devices on separate VLANs or separate wireless LANS and keep them firewalled off from your business network.
Jim Cashman
Senior Product Manager, Ipswitch

22. MULTI-SOURCE CLOUD MANAGEMENT

We tend to think of performance management as how people interact with applications, but IoT is about how things interact with things that interact with even more things, with a personalized human experience at the center of a tangled web. This new human experience creates a veritable ton of data in order to make our lives richer. For example, Oculus Story Studio is making virtual reality films you can interact with, Verily is working on a glucose-detecting contact lens for diabetes monitoring, and Moov one-ups fitness trackers by introducing real-time coaching to your workout. These devices create data that needs to be monitored and can introduces stress to the network. To handle this high-frequency data you need to introduce a distributed model for monitoring and embrace a multi-source cloud management strategy.
Shayne Higdon
President, Performance and Analytics, BMC Software

23. STANDARDIZATION

The lack of standardization for IoT protocols and security slows down communication for IoT devices, as messaging between the end devices and central infrastructure needs to be translated, encapsulated, encrypted, or all of the above multiple times along the path. Each time a message needs to be processed, this additional handling adds latency to the communication, degrading the IoT application performance.
Kimberley Parsons Trommler
Product Evangelist, Paessler AG

24. MACHINE-HUMAN COMMUNICATION

Many of us agree that IoT is one of the next big things which will have a significant impact on IT operations. Today, many focus on the sensor-device or sensor-monitor communications, protocols and security, but what happens when something wrong is detected or when the performance is deteriorating? In situations where human lives may be at risk or disasters imminent, I believe it is critical to think through the end-to-end use cases and pay special attention to the machine-human communication layer. Think about multi-modal notification, shift support, and escalation to ensure "someone" can take the appropriate action to fix the issue in a timely manner, should the issue be related to a pacemaker or a critical part of the engine of an airplane in air.
Vincent Geffray
Senior Director of Product Marketing, IT Alerting & IoT, Everbridge