NXP Semiconductors designs purpose-built rigorously tested technologies that enable devices to sense, think, connect and act intelligently to improve people’s daily lives. Cloud computing is a foundational technology for any NXP’s edge product portfolio as also its IT service. Their focus is on levering the public cloud.

Henk Coenen, Head, Cloud Center of Excellence at NXP: “We deliberately set up a Cloud Center of Excellence (CCoE) as a separate entity in NXP-IT, instead of embedding the CCoE in our IT infrastructure team. I’m reporting directly to our chief information officer”. In just over two years Henk Coenen has built his Cloud CCoE which provisions cloud services globally to +30 countries for +29k employees.

Prior to the initiation of the Cloud CoE in 2019, NXP transformed its application landscape drastically by migrating functionality from on-premise custom applications into Software as a Service (SaaS) applications. This journey started more than five years ago and focused on non-differentiating business processes, such as HR. The same is true for collaboration-related tooling. Henk Coenen: “As an example, we didn’t migrate our supply chain applications.

Furthermore, this strategy has paid off, especially during the pandemic, as we didn’t have any availability issues and our SaaS providers kept their applications compliant while continuing to make additional functionalities available. For non-differentiating business processes, SaaS applications are convenient and require minimal management attention from our Information Technology department.” Following the SaaS strategy NXP initiated the next phase of their cloud strategy, public cloud with the ambition of having 95% of the functionality in the cloud by 2026.

Due to compliance requirements and functionality related to real-time production systems NXP will never have 100% of its functionality in the cloud. The application owners are accountable for migrating to the cloud. As part of the Cloud CoE, a migration team supports the application owners. This migration has two phases: 1. a front-load phase to address topics like documentation and technical issues and 2. a migration phase. This enables an industrialized migration by NXP’s cloud migration factory with a quarterly migration cycle for the smaller applications.

Henk Coenen: “The first phase is the most difficult and complex phase. The cloud migration factory is very efficient for the smaller applications, we initiated separate programs for the migration of larger applications such as PLM and ERP.”

Phased approach

In the first year of the Cloud CoE phase, the focus has been in on detailing the cloud-first strategy, defining the policies & responsibilities, and addressing security challenges. In this first year also the cloud landing platform to deploy the cloud workloads/application has been built, following a proof of concept (PoC). “We selected a Research and Development design environment as the cloud workload for our PoC, which had a significant size in terms of both number of servers as well as data volumes, to prove that NXP is ready to successfully role out our public cloud strategy,” says Henk Coenen.

In the second year, the focus was on industrialization. NXP chose a phased approach, starting in Europe followed by the United States and Asia. Both Amazon Web Services (AWS) and Microsoft Azure have supported NXP in their journey. The current Cloud CoE is about 30 FTE and has teams in the Netherlands and India. “We are currently setting up an additional team to cover the North American time zones.”

Currently, NXP offers 2 operating models to support the cloud workloads running on top of the cloud landing platforms: Do It Yourself (DIY) cloud workload and a CloudOps service provided by Tata Consultancy Services (TCS). The cloud landing platform covers connectivity, Identity, and Access Management, security controls, and cost allocations mechanisms.

“By these two options, we are flexible and in control. At the end of the day, it is up to the application owner to decide which operating model works best for their applications. In the strategy of moving to the public cloud, we also rationalized our IT infrastructure services to one being TCS.” In the future, the DIY cloud workload will become more dominant.

Multi-cloud service stack

In their cloud strategy, NXP has selected both AWS and Microsoft (Azure) as their partners. A restricted multi-cloud strategy allows for flexibility while being centrally governed. Cloud providers have strengths, weaknesses, and unique capabilities, adopting a multi-cloud strategy allows NXP to access best-of-breed capabilities. A multi-cloud strategy allows NXP to quickly take advantage of new capabilities as soon as they are introduced by a vendor into the market.

Having a multi-cloud strategy implies that you have to think about the data flows between the Cloud workloads and the different cloud providers. Amongst other topics, this is driven by cost avoidance for cloud egress, as the current pricing models of cloud services providers are triggering charges for downloading data.

Henk Coenen: “The cloud workload owner is accountable for costs associated with its workload, as a Cloud CoE we advise them on how to architect their workload to optimize costs. During the front load phase in the migration, we often have to re-architect a functionality. This effort easily pays off.”

2030

The ambition is not to have 100% of the applications in the cloud. The focus of NXP will be on reducing the percentage of custom build functionalities. This is important not only from a cost perspective but also from a resourcing perspective. Finding the right talent is becoming more difficult. Talent is interested in directly adding value to the business.

“In that respect, the large cloud service providers, such as our partners AWS and Microsoft, seem to be a more attractive employer for IT talents to develop cloud services but in NXP IT talents will apply these cloud technologies to a cloud solution,” says Henk Coenen. Furthermore, the Cloud CoE will be positioned next to the business instead of between the business and IT. This will enable improved service provisioning to the business.