Looking back a year ago, Forrester Research predicted that cloud would “come of age” in 2019. Looking through this report, the predictions are largely accurate. Migration to cloud systems is accelerating, and business modernization did take off in 2019. This success had led to a mixed bag of deployment models including on-premises, hybrid cloud, and pure cloud platforms. The vision of a wholesale migration to cloud did not materialize.
The state of the market dictates that your cloud systems must integrate with applications and data sources that live on premises, in multiple clouds, or on hybrid architecture. To prevent these sources from becoming isolated silos, you need a holistic integration strategy based on forward-looking technologies. This applies to both application and data integration.
IT teams face a new challenge when trying to manage data across systems that span the options from on-premises to cloud. At Pega, we see it with our clients. They operate a mix of environments, and cloud migration considerations are focused on security and the changing data privacy regulatory landscape. These factors drive the need to operate systems using the right technology in the right environment. IT teams are rarely working in environments where management is consolidated across cloud, hybrid cloud, or on-premises – let alone a single cloud provider.
Cloud infrastructure providers are responding. AWS Outpost brings a cloud infrastructure stack into your datacenter, or Google’s Anthos allows companies to bring the cloud to their datacenter or manage multiple clouds from a single dashboard. These technology announcements reflect what we are seeing in the market. Systems will remain disparate. Successful IT strategies will have integration at their core to get the right data to the right system when needed.
Application integration in the cloud
The easiest path to application integration is to leverage a single platform with the flexibility to build applications that support multiple, related aspects of your business. When possible, this allows a single data store to secure and makes data available across systems. It would be nice if application integration was always this simple, but it’s not. In reality, you have to integrate applications that are potentially running on different technologies, using disparate data stores.
With the advent of the cloud, RESTful standards have emerged as the technology of choice for synchronous application integration. This change has been accelerated by the introduction of tools such as API (Application Programming Interface) gateways and developer tools. The advantage of gateways is that they make integration services easily discoverable and manageable. The drawback to RESTful services as a mission-critical backbone for your enterprise is that you need to develop mitigation strategies in case a key service is unavailable.
REST APIs are central to modern application integration. When considering your integration strategy, be wary of common integration terms that are positioned to confuse. For example, the term “Open API” is used frequently when talking about web-based integrations. When software companies use that term, it simply means they have made their API specification public so a developer may use it to integrate. Having an open API is just the first step. To better facilitate integration, most companies are adopting the Open API Standard (OAS). OAS provides a comprehensive human-and-machine-readable documentation standard for REST APIs.
Asynchronous messaging has seen a resurgence in popularity due to its resilient architecture and benefits for scalability. Messaging provides critical capabilities for IoT and is well-suited for high-throughput systems. While REST lends itself naturally to request-and-response integration patterns, messaging complements it by providing a solution for event-driven application scenarios.
This diversity in integration technologies and the fast-growing number of applications that need to be integrated has led to the emergence of integration-as-a-service (iPaaS) providers. These services provide a platform for developers to develop, maintain, and manage integrations to keep their fabric of applications connected when their application PaaS doesn’t provide the depth of integration options needed.
Data integration in the cloud
Application integration must be complemented with a well thought out plan for data integration. A successful data integration strategy is central to your cloud migration. Equal consideration should be given to security, data management, monitoring, and governance. Without understanding where your data is used and how it is housed, you end up with the same challenge you are trying to solve: data silos. With the right plan and governance, data can live in disparate systems and work seamlessly across applications. Integration can mitigate the lack of a “single source of truth” when one is not an option.
You also can’t really talk about data integration without talking about Big Data. Big Data is exactly what it sounds like: large data sets. Many of these datasets are brought together through integration technologies. In order to accomplish the objectives of Big Data processing, you need to bring together tools like data streaming, resource (compute) orchestration utilities, and massively scalable databases. Open source tools such as Apache Kafka and Cassandra have gained broad adoption in the market due to their ability to help get big data projects off the ground quickly.
Migration to the cloud has proven to be an evolution, not a revolution
More than 10 years after the launch of AWS E3 and S3, we are still talking about cloud migration. The wholesale migration of systems to the cloud hasn’t materialized; and most enterprises are actively using multiple cloud providers. Today, a cloud environment consists of on-premises, hybrid, and multiple public cloud providers. In this evolving landscape, the end-result of a successful integration strategy that accounts for security, compliance, management and governance in equal measures is an organization that is future-proof and able to combine new technologies with the accumulated intellectual property in existing systems.