Sorry, you need to enable JavaScript to visit this website.

Introduction to Fog Computing and its impact on Financial Services

Introduction to Fog Computing and its impact on Financial Services
December 15, 2017

Co-author:Ramakrishnan Kallidaikurichi

As devices that operate on the edge of the digital world get smarter and more powerful, they constantly capture data for analysis, aiming at gathering information on human and ecosystem behavior. Think of shopping habits to rainfall forecasts. At the other end of the spectrum is the Cloud. Seemingly, far away, there is a universe of computing and storage power for harnessing. Between these devices, connectivity and the cloud, emerges a perfect storm of convergence Internet of Things (IoT).

Gartner estimates that 30 billion to 50 billion devices will be connected by 2020.

The Fog Computing market worth is expected to be 205 Million USD by 2022.

Devices ranging from home appliances and personal wearables to industrial sensors will collect vast amounts of data. In today’s scenario, data is transmitted to centralized infrastructure (such as the Cloud) that can process it in a timely fashion to ultimately uncover insights and provide decision support. Additionally, there is an increasing demand to garner these insights instantly and continually adapt to the prevailing conditions. This means getting vast amounts of data (much of which could be noise) from billions of remote devices in a timely fashion to the compute hub (e.g., the Cloud), process it quickly and get the intelligence back to the device for auctioning.

Herein lies a conundrum – going back to Moore’s law. It is an unsustainable model if computing power on the Cloud is not commensurate with device and data explosion. However, along with posing the problem, Moore’s law also hints at a solution. Distribute affordable computing solutions when and where it is required. Go hybrid. This gives rise to the concept of fog computing – a term coined by Cisco. A simple way to understand this is to think of it as extending the Cloud closer to where the device and data source is. Just as in real life, while the clouds are up in the sky, fog forms near the ground, close to the action.

Fog computing helps address the top three demands of a digital-savvy world. To be successful, IoT, a fog cloud computing model must provide near real-time insights, address data privacy concerns, and enable affordability at scale. Cloud providers have started building regional infrastructure solutions to address some of these three concerns – but fragmentation of a cloud solution defeats its very purpose ultimately. There are edge computing solutions as well (utilizing the computing power on the device itself), but have the restriction of local device context only. The fog computing architecture brings together the best of both breeds by positioning itself between the Cloud and the device.

Fog nodes (compute points) are deployed at local area networks close to the device. These act as smart processors that can analyse data in the context of a community of devices in a localized domain, filter-out/pre-process and only send relevant data on the WAN to a Cloud backbone for further downstream analysis. These fog nodes are made available as hardware appliances with open platforms (Linux/JVMs) on which applications can be ported using virtualized programming and a deployment framework. Somewhat like complex event processing, data in flow is analysed/processed to draw inferences. Companies such as Flowthings, a start-up, have small footprint code that can be deployed on fog nodes, and new jobs are ‘wired’ together using a GUI interface. This allows developers to build new rules on the box and extend the same programming model to the cloud.

Cloud

IDC estimates that over 40% of the data collected by IoT devices will be processed closer to the edge. The fog computing market worth is expected to reach USD 205 million by 2022.

  1. Impact in FS space & Use Cases

    Fog cloud computing is a horizontal capability that complements both Cloud and edge computing. The theme is generally directed at tapping into services accessible from devices, such as payments, increasing personalized offers, and retaining customers who are lured by nontraditional banking services (Apple Pay, Samsung Pay, etc.)

    Some of the applicable scenarios include:

    1. Citi uses beacon technology enabling customers to use their smartphones for ATM access. Based on the location and the ATM users’ purchasing behavior, it provides incentives to use Citi cards in the local area.
    2. Access to real-time localized conditions give commodity traders a distinct advantage.
    3. Accessible solutions for smaller financial institutions who do not need large Cloud-based data crunching and insights, but rely more on timely and personal service to their customers.
    4. Real-time insights to insurance providers based on driving patterns, car conditions, traffic patterns at interchanges, and congested areas.
    5. Learning devices by leveraging fog computing-driven insights to improve the device itself. An example of this is drawing insights from family spending patterns to make all the devices in a local context (household) learn from each other.

    Financial services (FS) firms are increasingly looking to source innovation from IoT and analytics leaders in the software industry. Leaders such as Citi have thrown open challenges to crowd-sourced apps for IoT and analytics algorithms.

  2. Our point of view

    HCL is positioned to stay ahead of its counterparts in this space. To be effective in providing solutions, there are three fundamental capabilities a service provider should be endowed with – a) A robust partnership/alliance b) deep engineering (from embedded programming on devices to IP creation in analytical models and c) the DNA of innovation in the digital world. Our engineering capabilities exposed effectively using FS domain will be the clear winner. HCL is at an advantageous position precisely at the intersection of these three. Financial services at HCL needs to adopt a strategy to bring together these capabilities with a compelling offering to the IoT world –

    • Grassroots innovation for apps that leverage near real-time insights
    • Domain-specific, local context AI
    • Platforms to develop and remotely deploy, manage on the fog, extensible to the Cloud.

    Combined with the thrust given by HCL on Mode 2 (Next Generation Services), part of the Mode 1-2-3 for future proofing businesses, Fog Computing can propel HCL’s penetration into the IoT world by avoiding the pitfalls of now traditional cloud based approaches.

  3. References