Share this post:
I was recently invited to talk to a group of IT professionals on the theme of trends in cloud computing. Thinking about the brief and my experiences working in the field for the past few years, my talk centered on five major trends and signs of changes that I’m seeing on a day-to-day basis:
1. Advice to adoption
This is a trend I’ve already written about here. Essentially, enterprise focus has shifted from how to transform workloads to take advantage of cloud to how to truly adopt cloud throughout the business.
2. Low-hanging fruit to the upper branches
Cloud is no exception to the mantra “start simple.” Accordingly, many organizations have transformed their simplest workloads to the cloud and are reaping the benefits that cloud offers.
The next logical step is to tackle harder workloads with a lower cloud affinity, having learned from earlier achievements. Of course, that takes work. For example, some systems may need to be redeployed onto commodity hardware and operating systems from more niche platforms. This may require some application rework, but this is tending to come at a time when some kind of remediation work is required in any case.
3. DIY to SaaS
Not long ago, organizations were investing money in developing their own systems for every part of their business, but as businesses move their mindset from owning everything to using commodity services, there are three types of systems coming to the fore:
- Foundation systems needed to run any business, such as HR applications, accounting systems and word processing applications
- Competing Systems particular to an industry, such as supply chain applications, which are necessary to compete in the market
- Systems of innovation and differentiation needed to compete and win with customers; typically systems of engagement
Organizations are looking for pure software as a service (SaaS) for foundation systems. They want a mix of customizable SaaS and managed service using a cloud delivery model for competing systems. Internal development teams and partners such as IBM are concentrating on developing systems of innovation and differentiation, using platforms such as IBM Bluemix.
This allows enterprises to concentrate on delivering fast-to-market applications that meet the needs of their customers and reach new ones.
4. Monoliths to containers and microservices
In the pre-cloud era, the systems organizations implemented tended to be monoliths, doing lots of different and sometimes quite separate work. They’d often take a long time to implement, which carried the risk of not quite being fit for their intended purpose by the time they were up and running. Once they were in, they took a lot of maintaining, and a simple change could result in copious testing and a nerve-racking change release process.
Microservices developed using technology such as OpenWhisk are flipping that model on its head. Instead of having one single application that does everything, a microservice does one thing well. A business process might be completed by one or more microservices, and business services might share microservices too.
They’re typically between a few lines to a couple of hundred lines of code each, so writing them is much faster and maintenance is simplified. If a microservice is changed, only that changed service needs to be tested rather than the entire end-to-end process.
It’ll be interesting to see if the IT industry starts to monetize microservices and create a microservices economy in a similar vein to the API economy. After all, they operate on a pay-per-execution basis.
Similarly, organizations are looking to containers as a way to create application portability, mass scalability and, where appropriate, drive down costs through pay-per-execution models.
5. Single supplier to hybrid
A few short years ago, organisations were largely dedicated to a single supplier, but I’m now seeing them taking multi-cloud approach. This is in part because of a realization that not all clouds are equal and some workloads are better suited to different clouds.
This leads to some challenges, two of which are placement decision-making and direct connectivity between cloud providers to enable low-latency data transfer between applications hosted in different cloud environments. Decision-making around workload placement is being addressed by brokerage tools, IBM cloudMatrix being a notable example, while providing more direct connectivity between suppliers is problem that is being tackled at the organizational level for now.
The next major trend for my money is going to be in cognitive services, such as those which are already available from IBM Watson via IBM Bluemix. From intelligent chatbots to helping diagnose cancers and developing cures, developers will be enriching their applications with cognitive services so that they not only get a job done, but they can also support everyday decision-making and actions.
Learn more about IBM Bluemix.