Speed, Agility, Device: Choose any two
The Impact of Software Defined Everything on the Electronics Enterprise’s Agility
The challenge with device design: in general it takes too long. While software patches can be issued in days or weeks (or months if you’ve customized Android, see here), designing the next Xbox, refrigerator or IoT-based home security system is measured in years or large parts thereof.
Design of electronics hardware has always been the long pole in the tent from specification to user delivery. Very few have managed to collapse it down. From concept to research, circuit design and prototyping, OS customization, user experience design, testing and compliance, it’s a long road.
Yet, services design and the apps they bring with them is something that can happen rapidly and be updated with frequency. And users, most of whom don’t care a fig about the challenges and complexities, don’t like to wait and don’t understand why the latest functionality is only available to the newest buyers.
Software has eaten hardware’s lunch, and it’s not apologizing. And users — from consumers to Chief Operating Officers buying robotics for plants are all standing behind software saying “why can’t I have the expereince I want cheaper, faster and more accurately — on the devices/machines I have?”
The disruption software places on the hardware business drives two key impacts:
- The need to increase an organization’s speed and agility across the enterprise in multiple functions
- Challenges to legacy infrastructure, interoperability and trust, exacerbated across ecosystems
Getting from A to B or even A to X faster is more important than ever and requires some rethinking of how work gets done. Oddly enough, the old expression of “getting one’s head out of the clouds” is erroneous now, as hybrid cloud will be mission critical to forward momentum, as will edge computing (especially with 5G). Next, making smart decisions about what to process where, especially with IoT and IIoT throwing off tsunami waves of data has additional complexities — and requires rethinking the process architecture as well as the IT backbone.
The need for speed (and agility)
To put it in complex terms: “Virtualization causes hardware commoditization, disaggregating value chains and driving the need to increase an organization’s speed and agility.” If that seems like tech-speak, it is. I’ll break it down a little bit so even I can understand it.
Flexibility: The movement of computing and memory resources, including storage, and communication to cloud computing or edge allows as-needed scale up and scale down of processing power. That drives enormous cost reduction, processing speed and flexibility. It also enables things like “pay per use” business models.
Hardware virtualization or platform virtualization “creates” a virtual machine (VM) that acts like a real computer with an operating system. Software executed on these virtual machines is separated from the underlying hardware resources. — Peter Xu
With software, you can add on services that surround that need without any change to the hardware. As such, virtualization commoditizes devices — as small as a semiconductor or as large as a server, demanding new sources of differentiation in software and services.
Differentiation and Scalability: So these new software-defined device functions create new opportunities for differentiation and also the potential for faster, more fluid scalability and prolonged viability. By updating software, the need for new devices can diminish.
At the same time, this differentiation is favorable for chip manufacturers who see promise in AI-enabled designs, requiring specific high performance computing functionality. These fit for purpose chips address the balance between high performance computing and low wattage, with special workflows that address unique needs, especially at the edge.
Data Centers: Consolidation and innovation in data centers has affected some parts of the industry but overall allows greater options to meet the needs of data-heavy applications such as IoT, AI and analytics. Investments in cloud data centers continue to drive industry consolidation.
Ecosystems: Since IT departments can outsource this functionality rather than owning it, understanding capabilities, architectures and processes of the vendor community is vital. This allows IT (or marketing or product design) to focus on enabling core functions. At the same time, it means understanding and assessing security across a much broader scale. It also means designing and implementing policies around OT and well as IT. (3)
Challenges to legacy infrastructure, interoperability and trust are exacerbated across ecosystems — especially with IoT
IoT was embraced with blazing speed by many electronics companies both inside devices and across manufacturing. Connecting large numbers of previously-independent devices has unpredictable and often undesirable effects. And there’s still a considerable gap between being aware of the risks and mitigating them. Traditional infrastructure will be continuously challenged to manage this chaos. While security is of utmost concern, it’s not the only worry that might keep executives awake at night. Untangling the virtual cords that now connect devices to each other is complex.
Collaboration adds complexity: Communities of developers, spanning hardware, software, and data, face threats around interoperability, adjudication, data ownership and intellectual property protection. There is not a consistent standard that allows interoperability at the device, protocol and data layers. This lack of interoperability has even created a market opportunity for some companies like Unified InBox, whose UnificationEngine aims to solve that problem. The old standard, IFTTT, also works, even if its a little cludgy for the average user.
Latency: Edge computing streamlines the flow of traffic from IoT devices and provides real-time, context-specific local data analysis.(1) Increasingly, decisions will be made at the edge. You don’t want your autonomous vehicle pinging the cloud before it decides to move over for that cyclist on your right. Whether it’s traffic control systems, health devices or retail applications, more and more systems will reflect a combination of edge + cloud to address latency.
Data Integrity: The range of variability, veracity, volume, and velocity (4Vs!), and the differing quality and complexity of the data sources reduces value, sows confusion, and compromises security.
And even with all of these Vs, we need a new one to describe IoT data… which has two separate freshness dates. The first is in remediation. The second is in trend analysis.
Remediation activities are freshness-dependent — they require action. If there’s the opportunity to schedule preventative maintenance or to act when a group of sensors is showing a problem, this is not something that gets better with age. Taking action allows the problem to be addressed, whether it’s an approaching weather system or a machine belt. Trend analysis is fine over time as it allows more instances to make better predictions. (we talk more about combining data in this post).
Intellectual property: Expanded ecosystems with additional hardware, software and data layers enable electronics companies to change their intellectual property strategies to protect their investments and differentiate their products, adding value. (5) (See this post from Sue Hallen on new forms of IP)
In the next post, Scott Burnett will talk about how the disruption affect the consumer. We hope you’ll follow us for that.
Connect with the author, Cristene Gonzalez-Wertz
For additional information, please see:
2. From Navigating the Cloud Continuum Electronics companies implement hybrid to deliver innovation, January 2018
3. Internet of threats: Securing the Internet of Things for industrial and utility companies
5. Dealing with protecting IP, please see https://medium.com/the-future-of-electronics/intellectual-property-are-you-ready-to-race-ahead-with-new-forms-of-ip-d21fc7aad831
Also see: “Intelligent Connections”, IBM Institute for Business Value