Many organizations struggle to determine the best ways to modernize their internal and external business processes and applications. In this process, it can be challenging to determine the best path forward, especially when the options sometimes seem to conflict.
Here are some of the most common approaches, based on conventional wisdom, along with some emerging alternatives contrary to traditional thinking. You might be surprised at which approaches better fit your organization’s specific needs.
Conventional wisdom: Go all in on public cloud
Contrarian advice: Embrace the hybrid cloud
Perhaps your organization has had some false starts moving to the cloud or modernizing the application development lifecycle. You may have worked with consulting companies that made the use of a public cloud provider look easy, only to discover that your conglomeration of custom-developed enterprise systems, off-the-shelf CRM, and myriad third-party SaaS, PaaS, and IaaS don’t map cleanly to the services provided by the public cloud services you hoped would solve all of your problems.
The conventional wisdom has been to go all in on public cloud. But despite all the benefits of public cloud, it’s not a silver bullet.
[ Also read Hybrid cloud: 5 ways it can improve the customer experience. ]
The contrarian advice, which is now starting to enter the mainstream, is that it’s time to fully embrace the hybrid cloud. Companies are learning that while public cloud still has many benefits, the cost over time adds up. As an organization grows, there are usually opportunities to do at least some of the workload in the private cloud to gain benefits in locality, data transfer, and flexibility of in-house customizations.
Other considerations of the private cloud include various compliance, privacy, and security advantages. The hybrid cloud can offer “best of both worlds” benefits, such as edge computing and more effective paths to advanced technologies such as artificial intelligence and machine learning (AI/ML).
The “best of both worlds” effect comes into play when taking advantage of APIs and solutions that are open source and based on open standards. One example of this is running a workload that is simple to run in your private cloud and only has speedup when run on specialized hardware such as a supercomputer, quantum computer, or AI/ML (or other workload-specific) hardware.
Edge computing can be advantageous in situations such as manufacturing or transportation, where moving data and/or computation away from the edge device (e.g., factory machines or cars, planes, or other vehicles) is not feasible or practical with current technologies. In these cases, it is beneficial to bring the private cloud to the edge for domain-specific computation. Additionally, computation at the edge can provide privacy benefits because you can create rules around what data is transmitted or stored.
Another benefit of the hybrid cloud is the ability to move in the direction of specialized hardware after it has been proven out (perhaps even proven out in concept using the public cloud). Specialized hardware can be expensive and difficult to obtain. Working with companies that offer specialized hardware can be a great way to test the waters, but once you find the right hardware for your specific use case, bringing it in-house to avoid lock-in and subscription fees makes more sense. However, it can be useful to first test AI hardware or quantum hardware on the cloud to find the appropriate use cases.
Conventional wisdom: Reduce spending in uncertain times
Contrarian advice: Invest now in AI, research and development, science, automation, and future tech
A recent article by Andrew Ng posits that “the transformative value of AI is more financially powerful than interest rates.” AI is not the only option for getting the most return on investment during uncertain times; there are many benefits to investing in automation and platform engineering. There is also a lot of application for research and development in AI, the sciences, and any domain-specific area. Innovations that make an impact – for example, the development of cutting-edge software or hardware – can have a much greater return on investment than even the stock market.
[ Want best practices for AI workloads? Get the Ebook: Top considerations for building a production-ready AI/ML environment. ]
Another reason to take a “bull market” approach to innovation is that during uncertain times, geopolitical situations can change the landscape. Organizations of all different sizes may decide to leverage funding from both public and private sources. Examples include computer chip manufacturing and climate/green initiatives. Before you limit or halt your digital transformation or research and development efforts, consider the potential upside to investing big in areas that are now proving profitable. (Did I mention AI/ML?)
Conventional wisdom: Focus on commodity hardware
Contrarian advice: Embrace ASICs, AI hardware, custom domain-specific solutions, and edge computing
When making widgets, it’s important to create economies of scale to maximize profits. Public cloud providers have also relied on commodity hardware to scale and become profitable even in a competitive landscape.
One challenge for public cloud providers is to provide access to more specialized hardware at a reasonable cost. For example, even general-purpose GPUs are fairly expensive when rented in the public cloud. Over time, however, even GPUs will become commoditized. Remember that Moore’s Law has moved beyond packing more transistors onto chips and adding more CPU cores. To keep up with AI/ML data processing doubling every 3.4 months, the current generation of innovation relies on application-specific integrated circuits, AI hardware, and other custom domain-specific solutions.
The industry has only recently experienced this inflection point (away from commodity hardware solutions and towards domain-specific solutions), so there is a lot of opportunity for innovation. For example, many old and proven technologies, such as analog hardware, have become new again to reduce power consumption. This is particularly interesting to solve at the edge, where available power may be limited.
Wherever you are in your digital transformation, consider this contrarian advice and how you might apply it to your situation.
[ Discover how priorities are changing. Get the Harvard Business Review Analytic Services report: Maintaining momentum on digital transformation. ]
