Blogs

Based on the hands-on involvement of the Xona Partners team in the design and rollout of over half a dozen smart cities, the cascading technology trap emerged as a common thread: technologies are evolving at a significantly faster pace than the ability of cities to adopt.

This is not to diminish the importance of other challenges facing smart cities, such as financing, establishing optimal public private partnerships or handling of organizational structures within smart cities in a way that is consistent with governance models. However, for cities determined on implementing smart city applications, the cascading technology trap proved to be the primary and most complex challenge to overcome after having resolved other challenges.

The Cascading Technology Trap

Cities were determined at exploiting technology to improve operational efficiency and the life of their citizens, yet technology development was at the forefront of challenges facing these cities. The crux of the challenges is based on these factors:

1. Information Technology has been mainly driven by the leading Internet and Cloud companies (e.g. Google, Amazon, Microsoft) over the last decade. These companies have fundamentally different models of developing and deploying technologies, which made it difficult for the rest of the industry to absorb and adopt. Smart city organizations see an even more exasperated challenge in doing so.
2. The rapid evolution of Information Technology exceeds the ability of cities to assimilate knowledge, make decisions, plan, design and deploy a particular technology at scale. This creates strong competition between multiple technologies with competing ecosystems and little stability.
3. Technological leaps are not only moving at a rapid pace, they are also increasing in complexity requiring very sophisticated skills that come at high cost. In fact, commercial entities have been battling to acquire those rare skillsets, making it harder for government organizations of smart cities to achieve the same.
4. The organizational structure of cities, the decision-making cycle, the process of evaluation and deployment is too slow to assimilate complex modern technologies that cut across vertical silos around which city functions have developed.
5. Cities plan over the long term with expectations for mature technologies and validated business case. Modern technologies have a short lifespan relative to what cities seek. Often, the business case is not validated for wide-scale deployment. Validating the return on investment is a time-consuming activity.
6. Modern technologies are increasingly reliant on virtualized/cloud environments, which accentuate the need for specific skillset in software and programming languages that many cities either don’t possess, or find hard to attract.
7. Finally, standards have not evolved at the same pace as the technological advances, and it is difficult for risk-averse organizations to make bets on which technologies will win. This has also resulted in a large number of technologies competing in similar applications. In parallel, open source models have taken the lead in evolving technologies, making it even harder to bet on standards coming to fruition over a short-term horizon.

Examples of the Trap

I’ll give two examples of the technology trap in action. The first is the choice of the connectivity technology. The second is approach to the data layer architecture.

Connectivity technology options: A large number of connectivity technologies are available to connect devices used in smart city applications. Focusing on wireless technologies, we categorize them broadly into short or long range, and/or licensed or unlicensed spectrum. Examples include Zigbee, Bluetooth, Wi-Fi, LoRa, LTE-M, NB-IoT and many others. Analysis of how connectivity technologies have evolved over time leads us to two conclusions:

1. Wireless connectivity technologies are coming on market at a fast rate that is even a challenging for the ecosystem players themselves to decide on which technology to back. The challenge can only be amplified for the cities.
2. The proliferation and fragmentation of technologies each with different go-to-market strategy presents a complex mix of choices for cities to choose from.

 An example of this is the evolution of 3GPP cellular IoT technologies and their competition with others such as LoRaWAN, Sigfox, and RPMA.

The data layer architecture: A smart city data architecture integrates three distinct architectural layers: data source/sink layer, analytics layer and application management layer. Because of the breadth of this topic, I will give an example of the trap focusing on the analytics layer which is interface between the data source layer and application management layer.

The analytics layer deals with data storage and transformation to usable form, and with data processing through analysis models to churn out important insights, which are used by the application layer. The analytics layer includes a knowledge base, which has apriori domain knowledge rules, user profiles, and pattern information to help in data analysis. After data is transformed and aggregated, several rules from the knowledge base are applied to generate alerts in real time and outputs in forms that are easy to ingest by the application layer. Said in another way, the analytics layer is where the intelligence of the city lies and where complex decisions are processed and taken. This is the layer where the big data techniques and data sciences techniques come into play.

Data management techniques have been rapidly evolving making design goals a moving target for most organizations. Adding to the complexity of making the appropriate data management decisions at scale is the selection of the appropriate deployment model. Three such models have been popular over the last decade. First, the open source models where the leading cloud/internet companies open-sourced their big data followed by their artificial intelligence code. Some fast followers built open source experienced teams and led their own deployments. Second, a consolidation happened around the leading open source models, which forced a concentration of options for the end users. This has been the case for the big data solutions demonstrated by the offering of MapReduce, Cloudera, HortonWorks, Pivotal and a few others centred on Hadoop Big Data frameworks. A similar consolidation occurred for AI and Machine Learning around frameworks such as TensorFlow, Mahout, Torch, MLlib among a few others. Third, a parallel track emerged of commercial vendors with proprietary Big Data and AI implementation and niche market specialization. The variety of choices and constant shift in optimal cost vs. functionality outcome has made it challenging for many organizations, including smart city design teams, to make timely decisions.

To summarize across the data layer architecture, we see the following critical factors: 1. Technologies are being introduced at a fast rate and are especially bolstered by the open source approach. 2. New technologies rely on modern software development processes such as CICD. This stresses smart cities teams who need to possess rare skills to take advantage of these new technologies.

I have focused this article on explaining the cascading technology trap. So now that we understand what it is, and the challenges it brings about, how can we begin to address it? Stay tuned for a future post where I will share a few strategies that have been used by leading cities to overcome the trap.