Performance Issues of data generated by IoT

Atzori, Antonio and Morabito (2010) highlighted some of the high level issues related to the Internet of Things (IoT), which include standardization, addressing and networking, security, and privacy. There are several standards currently available and developing to assist the IoT, most important ones are namely, ZigBee, NFC, M2M and so on. Addressing is another important issue signified by the billions of IoT devices connected in next few years, which will require IPv6 instead of currently widely used scarce addressed based on IPv4. Networking protocols need evolution to accustom to small amount of data exchanges, which require modification in currently widely used transmission control protocol (TCP) based on three way handshake. TCP congestion control mechanism require evolution because wireless medium bring reliability and performance problems. Also for small transactions where session conclude with the transfer of single segment, TCP congestion control is not useful. TCP data buffering require buffering at both source and target ends, which is not cost and energy efficient for minimal to battery less devices. Security issues around IoT require careful resolution because most of the IoT devices remain unattended for extended amount of time but also traditional authentication and data integrity procedures are not suitable for small energy and cost limited IoT devices. Proxy attacks pose huge problems for IoT devices, because an illegitimate node can mimic identity of legitimate node and trigger response from IoT devices lead to loss of security (2010).

Pulipaka (2016) mentioned massive data aggregation process as another performance bottleneck due to high volume and verily structured data from IoT devices undergoing extract, transform and loading processes at servers. Indexing scheme is used to assist large scale web data generated by IoT however due to time, location and billions of devices, the observation and monitoring becomes highly complex. These indexes are embedded in IoT database applications for analysis, fusion, filtering, processing, and aggregation but high volume of real-time inserts becomes cost prohibitive on the performance of the system. Several indexing schemes like window indexing, dynamic indexing, time indexing, wave indexing and multi-granule indexing are used to balance cost and performance trade-offs (2016).

Deployment of IoT is easy however the data engineering pose significant performance bottlenecks, which include real-time processing, archive or prune the data, data sharing, handling huge volume, gathering data from all over the world, and most importantly moving prohibitively high amount of data (Woods, 2014). IoT likely require a data supply chain to accomplish all above tasks efficiently. Attunity is the technology built to move, replicate and synchronize complex data from IoT. Maestro works as the central coordinator for data supply chain by streaming data from large number of sources and aggregate at a central console to monitor movement of data in complex networks, akin to supply chain management of goods around the world. Attunity is successfully used at casinos and hotels to offer near real-time offers and analytics (Attunity, 2018). Verizon deployed Attunity and Hortonworks solutions for financial reporting and analytics in real-time (2018). Zebra is another technology developed by MIT, recently acquired by Motorola to provide end-to-end IoT solutions to analyze data in real time from workers in the field using cloud (Woods, 2014). ThingWorx is another technology works on graph technology by relating millions of devices interacting with each other and applications using Cassandra NoSQL database (2014).

Reference

Attunity (2018). Attunity. Retrieved from www.attunity.com

Woods, D. (2014). Data engineering is the bottleneck for the internet of things. Retrieved from https://www.forbes.com/sites/danwoods/2014/04/30/data-engineering-is-the-bottleneck-for-the-internet-of-things/#12a75b9a6247

Pulipaka, G. (2016). Resolving large-scale performance bottlenecks in IoT networks accessing big data. Retrieved from https://medium.com/@gp_pulipaka/resolving-large-scale-performance-bottlenecks-in-iot-networks-accessing-big-data-b0e386c58796

Atzori, L., Antonio Iera, A., & Morabito, G. (2010). The Internet of things: A survey. Computer Networks, 54(2). 787–2,805