Fluid computing for Supply Chain Management : Next economic disruption

Fluid computing is fast becoming a reality as the digitalization of services and products continues to grow. Right from self-driving cars to smart home appliances, every product is developed to utilize a cloud-based architecture to orchestrate them as a group of devices, applications, and sensors aligned to facilitate common objectives. However, due to bandwidth limitations and latency-based potential hazards, cloud computing cannot be used in a full-fledged system with intercoupled CIoT and IIoT systems.

Fluid computing is the umbrella term used for a distributed cloud, mist, fog, and edge computing system. Today, we will understand what Fluid Computing is and how it can help redefine the way supply chains function. Dive in deeper to explore the world of fluidic intelligence flowing across devices and applications.

Every futuristic city is imagined with self-driving cars, intelligent fleet management, smart infrastructure, health, and citizens. All of them need to be interconnected, and the current infrastructure cannot support such hyperconnectivity instances in a continuum. The cost of providing adequate bandwidth and latency would make such networks unaffordable and monopolize the internet data market. This will also bring back the era of telecom giants controlling the market with very little competition and overwhelming entry barriers.

I would also like to point out the fact that streaming a half an hour video content on an OTT platform like Netflix releases carbon emissions equivalent to driving a car for 200 meters. Even if we improve the current resource utilization rates at data centers, which are below 60% for CPUs and below 50% for memory, latency issues will still downplay the ecosystem.

The term fluid computing isn’t mainstream even in the tech industry. In fluid computing, the data flows across application states through replication and real-time synchronization by middleware. It allows the application state to flow in both synchronous and asynchronous modes to behave as a single system distributed among multiple devices with multi-person collaboration facilities.

As the name indicates, the cloud is the topmost layer, the mist is next in order followed by fog and ultimately the edge (devices.) The application state flows across all of the above-mentioned layers (server systems connected by a common network framework) and the edge to execute multiple processes in close coordination with all connected devices and stakeholders.

In simple words, fluid computing enables the localization of data processing. It allows you to push the intelligence towards the location where devices are being utilized.

For instance, if a wildfire breaks out in a smart city, the temperature control system of all smart houses will start alerting the firefighters. The smartwatches and all connected gadgets will alert users who will, in turn, try to leave using self-driving vehicles. Intelligent fleet management systems will try to divert the traffic while the parked vehicles will also start evacuating. The activity in such scenarios cannot be exactly predicted, and hence, it will pressurize the network beyond its capacity.

On the other hand, the bandwidth costs for meeting such requirements won’t be feasible for year-round deployment. It would be best if the devices process the data themselves to minimize their dependency on the network in such cases. Thus, to minimize the perils of cloud computing, a distributed networking infrastructure with autonomous decision-making capabilities for optimizing processing power with respect to network availability and the need for localization emerged in the form of fluid computing.

The eCommerce sector can be used as a classic example of how supply chains are facing the heat as the number of users increases. The cost of shipping products is increasing consistently as the number of deliveries continues to increase on a YoY basis. The more a parcel changes hands between pickup point and delivery location, the more it gets complicated, time-consuming, and costly. This applies equally to the entire global supply chain ecosystem, which is much larger, complex, and the demand-supply correlations govern it.

Connecting the manufacturing processes, sales forecasting, warehousing, order management, transport, and distribution for an integrated supply chain will require reinforcing all components with data processing capabilities. The integration of all stakeholders ranging from demand generators to demand fulfillers will require a mammoth network. This translates to connecting the CIoT systems of the demand generation centers to the IIoT systems of the manufacturers, logistics, distributors, and sellers to form a hyperconnected supply chain.

This will reduce the bottlenecks and allow the supply chain to function with unparalleled agility. This could help suppliers understand the demand patterns better and mobilize their distribution systems in a predictive way. Thus, it will lower the time required for moving goods through the supply chain, reduce costs, minimize wastage, and utilize the resources in an unprecedentedly optimized manner. In a way, fluid computing is going to become a breakthrough for supply chain management since it will focus on refining the exchange of goods across touchpoints instead of minimizing them through arbitrary compensations.

In this section, I am dividing the equipment and processes into two parts: demand generation and demand fulfillment. This will make it easy for readers to understand the application of fluid computing. Read ahead to see how various devices and processes will complete the supply chain:

Demand Centers

Fluid computing-based CIoT (Consumer Internet of Things) and software-managed processes form part of the demand generation part of the supply chain. The primary sources activating the supply chain will be the consumer devices. Home appliances like gas stoves, refrigerators, and deep freezers act as demand indicators for food, while the CIoT-based appliances for cooling, heating, ambiance, security, and housekeeping, along with utility services data, will be used for predictive maintenance. Apart from the consumer-initiated orders, the real-time data regarding demand consumption will make sales forecasting more accurate. Mostly, the consumer end will provide data for the supply chain management using sensors on their devices.

Sensors will be employed for identifying the current location of the demand for generating appliances and processes along with the stakeholders.
They will also help understand the current availability of goods at demand centers in terms of on-premise, under transient, pre-ordered, and stock available for sale at nearby distribution channels/centers.

They will help track the real-time shipping status of products and recommend consumption pattern changes as required.

They will also authenticate the demand figures to avoid hoarding and under the availability of products at the end-user level.
Sensors will also allow users to track the health of their devices and consumables used in them.

The important data regarding the rate of consumption, consumption patterns against historical demand data, and expected replenishment volume dates.

However, the appliances and CIoT networks will also require processing power to localize the decision-making. This means minimizing the synchronous data transfer for computation and shifting to asynchronous report exchanges heavily. If a device is suspecting breakdown, it should follow a standard operating procedure and prompt maintenance teams without waiting for centralized servers.

In the same way, the data transfer should materialize between the demand center and manufacturing/logistics firms without mediation. The role of centralized servers should be reduced to analytics and ensuring fair practices while doubling as a shared ledger for all stakeholders of the supply chain. They shall act as compliance enforcing systems that distribute data through multi-tenancy solutions to streamline order processing without damaging any stakeholder’s rights.

There will be a need for integrated shipping management systems that facilitate a common infrastructure for the delivery of all products towards the demand centers. Perhaps, the smart fleets of private and passenger vehicles will form the most important part of the supply chain in the future. Instead of shipping vendor-owned vehicles, using readily moving smart cars can help lower users’ transportation cost by allowing the logistics companies to utilize a certain portion of their storage spaces in their vehicles in exchange for compensation. Swappable storage compartments may also become a reality along with swappable batteries in the future. Robots will also play an important role in servicing the CIoT systems along with facilitating and authenticating doorstep delivery of products.

Demand Fulfillment

The demand fulfillment portion consists of manufacturing, distribution, and logistics mainly. This portion also includes e-commerce sellers and retailers who drive sales through their marketing activities and create demand with the help of advertisements. In the below section, the latter part of the supply chain is described:

The autonomous inspection systems will deploy heavy quality control for inspection of raw materials and semi-finished goods that will filter out any substandard item autonomously and send necessary reports to the concerned parties.

The warehouse management systems will also utilize sensors to track the storage conditions of the raw materials and finished goods within the premises.

Inventory management will become a distributed process carried by all stakeholders.
The purchase management solutions will track the goods in transient and prompt the user in case of any probable shortcomings. They will also keep secondary suppliers on hold in case of expected failure in fulfillment.

The orders made on the eCommerce sites and by retailers will be used along with user-generated data for better sales forecasting and streamline forecasting.

The manufacturing facilities will be heavily automated with the help of sensors, actuators, robots, software-backed processes, and quality management.

Asset management and resource utilization in IIoT-enabled manufacturing facilities will allow unparalleled transparency coupled with state-of-the-art security systems to lower manufacturing prices without compromising on product quality.
The smart fleets used for transportation and deliveries will use various sensors to track the condition of goods and vehicles. In case of any natural disaster or major blockage of roads in the travel route, they will also help in deciding the best alternative routes.

The logistics companies will be able to tie up with maintenance technicians and garages along with moteliers for minimizing breakdowns.

Using the fluid computing infrastructure on the demand fulfillment end will result in smarter individual devices, IIoT software packages for individual applications, and extreme localization at all cascading processing centers.

The resultant system will resemble an ecosystem of interconnected devices instead of infrastructure. The cost of using computation systems will surely increase initial costs for all devices and supporting systems.

The manufacturers of all appliances and machinery for both consumer and industrial users will have to upgrade their devices with Digital Signal Processors (DSP), Media Processors, Microcontrollers, Microprocessors, and Embedded Processors. This essentially translates to a complete makeover of all equipment designs and reversing the current approach of adding network capabilities instead of creating autonomous devices.

The roadmap for implementing fluid computing must start by establishing the equipment design standards for common utilization by third-party stakeholders through a unified architecture of middleware. This will require the device and appliances manufacturers to consider certain design changes and modifications that don’t fit into their product’s scope. Retrofitting could be a cost-effective solution for heavy machinery and other capital-intensive assets.

The fluid computing middleware should allow a great degree of freedom to connect devices to make decisions and interact with their deployment environment. A globally recognized software testing standard will also be needed to ensure the security and successful deployment of various devices of a different make. They should be similar in nature to the ISO standards for IoT software testing. They shall also include provisions for microservices-based solutions to incorporate better stability despite advances in different device technologies.

Consider the case of ordering a shaving product- Gel. Your razor will come up with sensors that not only track your shave quality and tell you if you’re missing on some spots of your beard. The razor will also order a replenishment using your shaving gel consumption rate and prices to purchase the right order quantity. Your shelf will also provide data regarding the current stock and storage space available. This way, only the required quantity is ordered for the best deal available.

The manufacturer of your preferred shaving gel will receive the order in advance before you empty your current stock. They’ll purchase only the required quantity of raw materials and at the right time. The shipping company will pick up your order and use smart fleets with swappable compartments to deliver it to your doorsteps. Chances are, your neighbor will receive your shaving gel at a traffic signal right in his car’s boot storage, and a robot will pick it up for you from his garage and place it on your shelf.

Thus, you can continue enjoying perfect shaves without even looking at how much shaving gel you’re left with, let alone driving to a nearby store to get it. Sounds great, doesn’t it?

Apart from pushing intelligence towards the edge, fluid computing will also develop distributed architecture models that act as interdependent entities that follow computation protocols for defined use cases and communicate efficiently without consuming huge network bandwidths. The telcos are also viewing 5G as their bid to bridge the networking and infrastructure gap between consumers and industries, finds Capgemini. We can also expect its collaboration with other advanced technologies like machine learning, AI, and blockchain. Fluid computing will surely cause the next economic disruption by expediting the flow of information and goods at an unprecedented speed, transparency levels, and at reduced operation costs.