JNT Visual - Fotolia
Software-defined infrastructure (SDI) is set to become mainstream in 2018, according to analyst Forrester.
SDI, in conjunction with automation and the use of bots in the software development lifecycle, will help IT to keep up with the rapid pace of change demanded by the business, it said.
The analyst firm’s Predictions 2018 report said: “Thanks to increased investment by supplies, as well as rapidly maturing infrastructure automation toolsets to manage them, uptake of software-defined infrastructure has never been faster.”
Forrester said many of its clients have looked beyond proof-of-concept deployments of SDI to see how to deploy critical workloads. “In 2018, a strong inflection in software-defined infrastructure of many sorts will fuel a fervent community adopting infrastructure-as-code philosophies and practices,” it noted.
Unless there is a real technical reason not to, Forrester recommended that production workloads should be placed on software-defined storage and compute platforms. But it warned that software-defined networking is still less mature.
The analyst recommended IT departments to use configuration management and continuous delivery release automation software to ensure consistent infrastructure models across development, test and production.
CIOs should also encourage their teams to explore Azure Stack, VMware on Amazon Web Services (AWS) and similar technologies to link on-premise workloads with the public cloud, said Forrester.
As the pace of change continues, IT departments will embrace continuous deployment, the analyst said. It pointed out that 65% of organisations have shifted from yearly and half-yearly release cycles to monthly or faster and, as the velocity of application development continues to increase, IT teams will ultimately give delivery updates on an hourly basis.
Read more about software development best practices
This will lead to increased use of automation, which will be used to deploy applications, middleware, database and infrastructure quickly and consistently, said Forrester. Its report added: “Many large organisations have already invested in continuous delivery release automation [CDRA] to drive the deployment of applications within production environments. While these tools have the capabilities to deliver complete releases throughout the pipeline, many organisations have not yet taken full advantage of this.”
As code is released more quickly, Forrester predicted that software testers will turn to bots to help them automate testing procedures. “We predict that testing teams under the pressure of agile and continuous integration/continuous delivery [CI/CD] will be leading the automation race in the software development lifecycle and position themselves to crack some hard problems of testing AI solutions, too,” it said.
Forrester recommended IT teams to use bots at the start of their software development processes to help developers and testers optimise and validate requirements and test cases by scanning them with natural language processing [NLP]. Forrester said bots could also be used to classify defects in code.