Castro added that StatefulSets gives traditional applications “a place to go,” so users can reap the benefits of Kubernetes without being forced to rewrite applications.Kubernetes 1.5 sort of marks a passing over the hump of “It’s interesting technology; we should take a look at it,” to, “We’re actively testing our applications on it,” said Castro. In version 1.6, companies should expect to see new opportunities for the community to scale out its Kubernetes expertise and knowledge, which in turn leads to more contributors and participants, he said.Another feature worth highlighting in this release is Federation, which means someone can take an individual Kubernetes environment and pair it or federate it with a second or multiple other Kubernetes environments to make them look like one pool of resources, said Gracely. He said this is important because companies can enable their data centers to use additional public cloud resources, allowing them to have a technology that allows them to think about how to go beyond the walls of their data center, but still have the visibility and control over what those environments look like.Limitations of Kubernetes 1.5StatefulSets, kubeadm and other features are still in beta, but they are seeing quick development, said Castro. The parts that still need work included installation, upgrades and configuration, because they are not where they should be, especially on bare metal, he said.“Every organization is different, especially networking, so it can be challenging to help someone set up a cluster if they don’t have a high level of control of their infrastructure,” said Castro. “We strongly believe that the only way to do this is to stop treating bare metal as a unique snowflake and to model it like a cloud. Users who use metal-as-a-service to model their hardware and network as a cloud have a significantly easier time standing up Kubernetes.”The one big thing Kubernetes will need to tackle is upgrades, because “releasing every three months is brutal for users,” said Castro.“Nailing a reliable and repeatable upgrade process is the thing I’m most excited about in 2017,” he said.Kubernetes community velocityKubernetes 1.5 was also evidence that its ecosystem is growing rapidly, which benefits partners like Canonical and Red Hat, which already provide commercial services for Kubernetes users.Castro said Canonical stands up and operates Kubernetes for customers through a program called BootStack, which is similar to what Canonical has been doing with OpenStack for years. Canonical is also standing up a production Kubernetes cluster for its snap store, which will serve software packages to Arch, Fedora, Gentoo and Ubuntu users, said Castro.Red Hat uses Kubernetes technology as the core of its OpenShift technology, and it delivers that as a software element to customers who want to run their own software. Red Hat also uses Kubernetes for its day-to-day operations for what it calls OpenShift Online or OpenShift Dedicated, said Gracely. This means Red Hat operates the technology for customers who just want to use containers and build new apps.Besides checking out Kubernetes partners for additional resources, Gracely suggested Kubernetes users read about recently added features and updates of Kubernetes 1.5, and to have teams “get their hands dirty” by testing it out. More companies are adopting container technology to achieve things like developer efficiency, and while Docker remains a leader in container technology, Kubernetes is a choice for container orchestration and management, according to a recent ClusterHQ report. The most recent version of the open-source project, Kubernetes 1.5, includes several new features that can be beneficial and help teams optimize container usage in their organization.The current state of containers has hit a tipping point, according to Red Hat’s director of product strategy, Brian Gracely, and businesses are finally saying that they need to be faster and more responsive as to what is going on in the market, he noted.(Related: Kubernetes gets support for Windows Server 2016)As companies try to iterate more quickly, they end up having to rely on individuals to update or make sure the systems are secure and scaling appropriately. Instead of doing these processes manually, companies can utilize Kubernetes so it run things automatically, said Gracely.“[Kubernetes has] become this technology that we were able to leverage out of the big global web-scale cloud providers, and it’s becoming very applicable to lots of vertical businesses and lots of vertical industries,” he said.When Kubernetes first came out, it was designed for the mindset of “Let’s use this technology to build new apps,” whether it be mobile or IoT or any other type of application, said Gracely. Now, the technology has created two paths for companies to take, where they can either apply the technology to new projects or figure out how they can add it to their portfolios, he said.“The latest release of Kubernetes 1.5 is really the beginning of the Kubernetes community saying, ‘We have something really good, we are seeing a lot of adoption, we are seeing a lot of contribution, and we are seeing real customers say help us solve more problems,’ ” said Gracely.Key features of Kubernetes 1.5Kubernetes 1.5 introduced solutions for those looking to run a distributed database on Kubernetes, including solutions on how to guarantee application disruption service-level objectives (SLO) for stateful and stateless apps, according to the Kubernetes team. Kubernetes 1.5 is also special because StatefulSet and PodDisruptionBudget moved into beta. These features “provide an easier way to deploy and scale stateful applications, and make it possible to perform cluster operations like node upgrade without violating application disruption SLOs,” the team wrote.Gracely said that StatefulSet allows companies to bring their applications right into a Kubernetes environment and run them in containers. The system gives teams much of the look and feel they had in their environments, but for customers, this means they have a more consistent platform and they do not lose out on functionality, he said.Jorge Castro, cloud community liaison on the Canonical cloud developer operations team, said StatefulSets are “the most interesting” feature of the 1.5 release, mostly because it “allows traditional workloads to have a place in the new world.”“It can be frustrating to hear about all the great things containers can provide you, and then finish that off with ‘…but you need to rewrite all your applications,’ ” he said. “That’s not an option for a large swath of enterprise applications. If I’m looking at cloud-native infrastructure for all my future applications, and I have still have to maintain infrastructure for traditional applications, now I have two problems.”
“The first stable release is an important milestone for the Apache Beam community,” said Davor Bonaci, vice president of Apache Beam. “This is a statement from the community that it intends to maintain API stability with all releases for the foreseeable future, making Beam suitable for enterprise deployment.”Beam’s data processing pipelines and runners can be executed on Apache Apex, Flink, SPark and Google Cloud Dataflow as well as other execution engines. Other features include a single programming model for batch and streaming use cases, ability to executive pipelines on multiple environments, and the ability to write and share new SDKs, IO connectors and transformation libraries. Apache Beam is already being used by companies like Google Cloud, PayPal and Talend.“We congratulate the Apache Beam community for reaching the key milestone of a first stable release,” said William Vambenepe, lead product manager for Big Data at Google Cloud. “We look forward to our Google Cloud Dataflow customers taking full advantage of Beam’s powerful programming model and newest features to run their data processing pipelines on Google Cloud.” The Apache Beam team has announced the first stable release of the project, version 2.0. Apache Beam is an advanced unified programming model designed for batch and streaming data processing. In addition to this being the first stable release of the project, the team says this is also the third most important milestone for the community. Apache Beam first entered the Apache Incubator in February 2016, and later became a top level project in December. The release features improved user experience, seamless portability across execution environments, API stability, stateful data processing paradigms, support for user-extensible file systems, and a metrics subsystem.
Prior to the digital age, businesses tended to define themselves according to their products or services. However, digital technology has fundamentally altered buyer expectations and habits in both the B2B and B2C sectors. Survival in the modern marketplace necessitates viewing customers, rather than the product or service, as the center of the enterprise, and embracing digital transformation to meet customer demands. Most organizations realize this: 83 percent of respondents to Forrester’s Q1 2018 Digital Process Automation Survey stated that they were going to expand the use of process improvement methodologies for customer journey-mapping within the next 24 months.Organizations must develop five to 10 times more applications than they did in the past. Meanwhile, there is a severe shortage of qualified software developers to meet the insatiable demand for apps. A survey by CapGemini and LinkedIn found that 54 percent of global organizations have been impacted by the IT skills shortage and forced to alter their digital transformation plans.Everyone agrees that to cope with this problem, organizations need to “future-proof” their digital transformation efforts by abandoning traditional development processes – which result in hard-to-maintain legacy code that does not scale – and find a more sustainable way to develop applications. Low-code development platforms were created to address this need. These platforms allow non-programmers who have strong general computer skills to quickly and easily develop and deploy applications using drag-and-drop modules, without having to write a single line of code. This greatly reduces the need for skilled software developers and allows companies to develop and implement applications much more quickly.However, sustainable, “future-proof” digital transformation isn’t just about implementing certain technologies; it alters the way in which organizations conduct business and helps support the organization’s strategy. Following are four requirements for a sustainable digital transformation.Do not focus solely on “fast and easy” application development.Arguably, the biggest myth of sustainable digital transformation is that it’s all about “fast and easy” application development; this is why “fast and easy” is a major selling point for many low-code software platforms. Future-proof digital transformation isn’t just about “fast and easy” app development. It is more about creating sustainable apps that improve and automate processes, easily scale across the enterprise, and can adapt to tomorrow’s technological and business needs. Do not think that a low-code platform alone is enough.Nearly two-thirds of respondents to the Forrester Q1 2018 Digital Process Automation Survey stated that process was a critical part of their business strategy, underpinning initiatives such as digital transformation and improved customer experiences. DPA is fundamental to ensuring that back-end processes align to support the on-demand, personalized customer experience modern buyers demand.Do not think that a sophisticated business process management system (BPMS) alone is enough.Most software vendors, including many low-code vendors, default to the “fast and easy” myth; they describe the biggest challenge of digital transformation as the ability to create more business apps faster and easier. However, in a customer-centric digital transformation paradigm, processes are the key to creating a predictable, consistent, and continuously improvable customer experience. Thus, the ability to create more and better processes is critical success to sustainable digital transformation. Sophisticated BPMS solutions are too complex and too costly to quickly the create hundreds, even thousands of mission-critical processes that organizations need.In customer-centric digital transformation, back-end “customer journey-mapping” processes are just as mission-critical as traditional back-office core processes (as well as front-end design), and they demand robustness and scalability. Organizations often use sophisticated BPM solutions to create their back-end business processes, with the key objective to optimize them for efficiency (e.g., Six Sigma/lean principles). They assume that their BPM solution can be used for customer journey-mapping processes as well. However, there are two problems with this:The number of back-end businesses processes required ranges from single digits to a few dozen. Customer journey mapping processes will number in the hundreds to thousands.The “end-to-end” nature of back-end business processes typically span different internal LOB systems and internal functional areas. In contrast, the “end-to-end” nature of the customer journey-mapping processes must work anywhere where customers might be; in addition to internal systems, these processes will also interact with external channels such as mobile and cloud services, IoT devices, and AI-enabled services.Due to the rapid evolution of technology, future-proofing is critical to the customer journey-mapping processes that support sustainable digital transformation. These processes must be dynamic at runtime (or hyper-agile) to deal with exceptions and seamlessly bridge today’s demanding digital customers from their current experience to a new experience.Jesse Shiah is CEO and co-founder of AgilePoint, a software development firm with a platform for building digital, automation apps quickly and easily while ensuring future-proof simplicity.