XOps and industrializing the technologies of the future

Increasingly, new technologies are coming over the horizon that don’t fit our standard approaches to build, deploy and support systems. Sometimes we lack the specialist expertise required to do so. We need to be proactive and better prepared for these changes, rather than treating them as an afterthought.

At Hybrid Intelligence, a Capgemini Engineering team, we conduct regular internal research projects to increase our capabilities in emerging technologies. One of our XOps teams has just completed a research project specifically exploring the challenges of bringing next generation technologies into production. As a testbed, we used current emerging technologies such as Digital Twin, Knowledge Graphs, EdgeML, and Quantum Computing. We examined the aspects of scaling, deployment and maintenance that lead to successful industrialization.

We created frameworks for these technologies enabling us to standardize delivery, improve turnaround speed, and increase the quality of the solutions we bring to our clients, thus allowing them to adopt these nascent technologies to accelerate innovation. However, more important than those individual frameworks, was that this project identified the common factors across all these different technologies, and thus informs how we should evolve our XOps team to be ready for next generation solutions.

Here’s what we learned:

Building future XOps capability

To integrate new capabilities within our XOps skillset, we started with some research into the technologies themselves, and utilized some of our XOps principles covering other areas to consider how they may apply to the nascent technology. We also referred to various quality frameworks such as ISO- 25010 (software product quality), ISO-25012 (data quality), CRISP-DM/ML, and more.

Using this information, we formed various questions such as –

• What does continuous integration and delivery of a quantum circuit actually look like?
• What are the challenges in maintaining and changing knowledge graphs and their associated ontologies?
• What are the key considerations that influence the quality of a digital twin?

At Capgemini Engineering we do have specialists in each of the subject areas, for example we have an internal Quantum Lab dedicated to advancing the state of quantum computing research. So, once we had formed our questions, together with our experts we explored the best way to industrialize and then operate solutions built upon these technologies. In collaboration with our experts, we then built operations processes and frameworks to formalize best practices.

Compare this to a non-XOps approach –a business utilizes specialists to build quantum PoCs and then to tries to bring that PoC to production, but without due consideration for the challenges present in industrializing solutions built on this technology. This came to be known as “PoC Hell,” and my team witnessed it early on as organizations sought to adopt machine learning approaches.

In doing all of this, we have built capability in our operations team to extend our XOps capabilities to new emerging technologies. We continue to make these kinds of research investments to extend our XOps capability further through the creation and industrialization of internal proof of concepts, the development of staff skills, and the collaboration between our operations and delivery teams.

Conclusion

Solutions are becoming increasingly complex, and often the greatest gains come from the convergence of multiple nascent technologies. XOps is the practice that will manage these solutions. We are committed to not only ensuring that we have the capability to deliver XOps now, but we are investing to ensure that we expand on our capabilities as future technologies mature.

We have made sure that, as businesses begin to utilize these technologies to build new solutions, we are ready to ensure that these solutions are robust, secure, scalable, and trustworthy.