App development teams brace for big change in 2018

PaaS platforms evolve

In 2018, vendors will attempt to improve their platform as a service (PaaS) offerings to speed up application development and support DevOps workflows. Leading cloud providers, such as Amazon Web Services (AWS), Google, Microsoft, IBM and Oracle, will continue to roll out tooling to automatically provision PaaS-like capabilities via containers and Kubernetes. In addition, CIOs and app development teams will start to take advantage of higher-level service mesh tools, such as Istio and Linkerd, to recreate the benefits of monolithic apps within a collection of microservices.

However, these more sophisticated cloud tools come with potential vendor lock-in risks, said David Bartoletti, VP and principal analyst at Forrester. For example, AWS provides a set of tools to automatically provision containers with its Elastic Container Service, but those tools won't necessarily help developers who need to move apps to private infrastructure or manage apps that span cloud platforms.

Still, both PaaS adoption and container-based development will continue to grow.

"The rush is on for enterprises to build and deploy better software faster, and that's going to drive a doubling of PaaS adoption -- both on premises and in the cloud -- in the next 18 months," Bartoletti said. "In some industries, like financial services and retail, leaders are already differentiating by how well they release high-quality experiences, and many of them are using a Cloud Foundry- or Kubernetes-based container development platform to speed up even further."

Growth of hybrid, multi-cloud apps

Interest in big data and AI continues to grow in the enterprise, but for these applications to succeed, they need to integrate data from multiple, distributed sources, said James Staten, global head of market development at Equinix and former chief strategist of Microsoft's cloud division.

As a result, more enterprises will distribute their applications across multiple cloud platforms, and those who already do realize benefits, such as workload deployment, will move to the most optimal environment based on factors like performance and cost. In 2018, there will be significant progress around technologies that can fully integrate multiple clouds to satisfy required application service-level agreements.

The rush is on for enterprises to build and deploy better software faster, and that's going to drive a doubling of PaaS adoption -- both on premises and in the cloud -- in the next 18 months.

However, as information crosses cloud platform boundaries, app development teams need to be wary of data leakage and potential violation of new privacy mandates, such as the General Data Protection Regulation. This is especially true in heavily regulated industries, such as financial services and healthcare.

Teams need to carefully orchestrate hybrid and multi-cloud workloads to address these security, data privacy and compliance challenges.

Blockchain gains enterprise foothold -- slowly

Blockchain technologies continue to grab enterprises' attention, with the rapid growth of new cryptocurrencies. In the coming year, app development teams will start to experiment with blockchain for transactions in financial services, manufacturing supply chains and secure data exchange. Major enterprise vendors, like Oracle and IBM, already offer blockchain cloud services that enable organizations to experiment with business processes rather than having to code the underlying blockchain logic.

In the coming year, these implementations will remain mostly experimental. Despite the excitement, blockchain transactions add a lot of computational overhead for most business processes. While proponents tout better security and reliability, some blockchain networks introduce new vulnerabilities, as demonstrated by attacks on cryptocurrency wallets.

Furthermore, blockchain mechanisms, such as smart contracts, introduce risks and challenges around rolling back fraudulent and unintended transactions. While safeguards for these risks are already built into existing transaction processing architectures, it will take time to figure out how to do so for some blockchain technologies.