An important announcement regarding the SHARE Virtual 2020 event: Due to unforeseen technical difficulties, the SHARE Virtual 2020 event will resume September 22-24 and September 29-October 1. Specific dates and times will be announced in the coming weeks. This article was published prior to the event reschedule and session/speaker information is subject to change. For more information, please visit the SHARE Virtual FAQ page.
Next week, SHARE Virtual will not only introduce members to the virtual platform Pathable when it kicks off Aug. 4, but the event will also highlight new tools and updated languages that attendees can use at their own offices to improve efficiency and productivity. Through technical sessions, networking events, an on-demand library, and other opportunities throughout the month of August, attendees will have an opportunity to learn how to adopt a mainframe-inclusive DevOps strategy, integrate batch scheduling and automation through Zowe, how ZIGI can help programmers get the most out of Git on z/OS, how the latest version of CICS can further empower developers, and how to use machine learning to predict bad things before they happen. We spoke with some of the session leaders to highlight what attendees can expect at SHARE Virtual 2020.
Madesh Ramamoorthy, DevOps – technology architect at Infosys, will host the “Multiple Platforms, One Tool Set – A Mainframes Inclusive DevOps Strategy for Enterprise DevOps Transformation” session along with his colleagues Elango Paulraj and Nirmalskandan Natarajan. In the session, he will explain how DevOps should be commonly used to align the people and products toward a mutual company goal. However, Ramamoorthy explains that such a shift in strategy needs to be driven by leadership and guided by a clear understanding of the mainframe landscape in order to be effective. He adds that the strategy will help companies reduce their technology footprint, demonstrate how to use the IBM tool stack to increase mainframe agility, and reduce the time to market for mainframe applications.
During his session, “How to Integrate Batch Scheduling and Automation in Your Hybrid Cloud Applications Through Zowe,” Domenico D'Alterio, offering manager for Z automation at IBM, will explore how Zowe can be used with z/OS to integrate batch scheduling and automation into hybrid cloud operations. According to Forrester, 72% of customer-facing applications are completely or very dependent on mainframe processing. D’Alterio says Zowe can capitalize on that through these modern interfaces to interact with z/OS and allow users to work with z/OS in a way that’s familiar to what they experience on cloud platforms. “The add-on of IBM Z Workload Scheduler for Zowe CLI (Command Line Interface) allows an easy integration of scheduling in DevOps scenarios, giving the possibility to take actions and monitor the batch automation and scheduling through commands, easy to integrate in scripts,” says D’Alterio. “Instead of creating a standalone CLI, we decided to offer workload automation commands in a standard framework like Zowe. It enables the end user to have workload automation commands in the same environment of other commands to interact with z/OS, JES, and more.”
For instance, D’Alterio says, “A Turkish bank leveraged IBM Z Workload Scheduler (IZWS) REST API in the homegrown Web UI provided to their developers for other internal purposes to make it easy to ‘submit a job request’ providing minimal information. IZWS REST APIs were leveraged to add on-demand jobs to the scheduling plan, submit the job added, and to retrieve job status and log once completed.” He adds, “Another example would be a Swedish bank leveraging REST API for IZWS as the fastest and easiest way to accomplish a third-party application integration with IZWS.” D’Alterio explains that when using Zowe in this way, there is a reduced learning curve because of its easy-to-use and “cloud-like” interfaces. Users will have the ability to include workload automation actions in scripts and applications, leveraging a standard framework (Zowe) to interact with z/OS, MVS, and JES, and easily integrate workload automation with JES, MVS, and z/OS actions through Zowe API Mediation Layer and CLI.
In the “Git for ISPF – Finally a z/OS Based User Interface” session, ZIGI developers Henri Kuiper, CEO at zdevops B.V., and Lionel B. Dyck, software developer at 21st Century Software, will explore their open-source ISPF application ZIGI. They also will demonstrate how it can help users leverage the capabilities of Git in an ISPF environment. ZIGI enables ISPF developers to manage their source code, load modules, or other non-VSAM data sets in a Git repository without leaving their trusted ISPF environment. One of the main drawbacks of Git on z/OS for the ISPF developer is that it operates on USS files and folders, where the ISPF developer often works in data sets and partitioned data set members. Without ZIGI, developers working on ISPF have to copy things from z/OS data sets to USS files and then get into the command line interface (within USS) to manage their repository. “ZIGI removes all these barriers and presents the ISPF developer with a pure ISPF application,” says Kuiper. Another problem is that developers have to know all of the detailed Git commands to push, pull, merge, and (sometimes) resolve merge conflicts, but he says that ZIGI presents the features of Git via the well-known ISPF interface. ZIGI allows for the creation of local repositories with local z/OS data sets and cloning of remote repositories, which Kuiper says improves how ISPF developers collaborate and share their tools and products without learning all of the Git commands.
William Yates, CICS test architect at IBM, says the “Relevancy of CICS in Chapter 2 of the Journey to Cloud (and What's the Latest for CICS)” session will highlight how the latest CICS release, CICS TS 5.6, can empower developers to help companies evolve applications and improve processes at a lower cost. CICS TS 5.6 provides Spring Boot, Maven, and Gradle support, as well as offers enhancements in security and policies to make CICS as robust as ever while providing more automation. “With well over 500 ‘Request For Enhancements’ (RFEs) satisfied in version 5, the foundation of CICS is core to the platform, with focus on resilience in this release, primarily in the areas of security and policies,” he says. Moreover, Yates adds, “CICS Transaction Server delivers capabilities that allow enterprises to take advantage of cloud native DevOps tooling and CI/CD, modernizing applications to meet the needs of today's development and operations teams, as well as maintaining the security, availability, and scalability that z/OS are renowned for.”
In “Predict Bad Things Before They Happen ... Even if You Haven’t Seen Them Before” session, Nigel Slinger, architect at BMC Software, will talk about how machine learning (ML) can understand a system’s normal operations in order to predict performance slowdowns, disruptions, and other operational hiccups. “Performance slowdowns manifest themselves in many ways so multivariate techniques that combine many metrics allow different views to be combined,” he says. “For example, more CPU may be needed to process the same ‘normal’ amount of transactions. More contention in the system may cause extra tasks to be created because the existing tasks haven’t given an answer.” Slinger adds, “Machine learning can give an indication of both under ‘normal’ and above ‘normal.’ Using these multivariate factors, the answer can be figured out.” However, he says that machine learning is dependent upon accurate models that rely on adequate representative training data. “Models have to be evaluated for accuracy based on their predictive answers,” Slinger explains. “If the models give the wrong answer, then the training data, feature engineering, and ML algorithm has to be validated. Changing and improving the multivariate feature engineering may well correct the issue. Good feature engineering is down to a key partnership between data scientists and domain experts.” Firms need to be concerned about false positives/invalid alarms, the overhead costs of training and scoring, and the complexity of the underlying ML infrastructure, particularly hardware and software requirements, he adds.
Meanwhile, the “Employing Multifactor Authentication in a Z Environment” session from Steve Kiernan, vice president of mainframe architecture and solutions at State Street Corp., will be available through the on-demand library during SHARE Virtual. He says the session will focus on how multifactor authentication (MFA) on the mainframe can help provide businesses with security confidence and shore up their systems against internal threats. “MFA significantly improves the confidence that any given user ID is, in fact, being used by the person it was issued to,” says Kiernan. The session will provide an overview of which applications and products don’t work well with passphrases, as well as those that employ password caching and replay. He says that organizations may have to engage in some research and development, if they have not already implemented passphrases and will have to address certain considerations and policy changes if they use Out of Band (OOB) authentication. “Since this is a change to the end-user experience, communication is paramount,” explains Kiernan. “You will need to create an ‘MFA-enablement Package’ that includes what is happening and why, how is the end-user experience changing, rollback if necessary, FAQs, and more.” Firms also need to determine if the people that need MFA enabled have the necessary authentication factors, including IBM TouchTokens and RSA hard and soft tokens, among others.
Through SHARE Virtual, mainframe professionals can learn about the latest tools and updated languages to improve efficiency and productivity at their own businesses. The virtual event promises to have the quality sessions members look forward to and much more. Throughout the month of August, expect to learn more and connect with others in innovative ways.