devopsy.kinsta.cloud

Tag: agile architecture

  • Crafting an Agile Enterprise Architecture

    Crafting an Agile Enterprise Architecture

    Digital transformation, business agility and enterprise architecture are frequently heard terms in this era of digital technologies and rapid change. The appropriation of digitalized applications, solutions and processes has enabled organizations to increase efficiencies. In turn, this allows them to respond to shifting customer needs and trends in a more flexible and agile manner. 

    Digital transformation, the rise in mobility and a remote workforce have created a shift in the business model paradigm. The adoption of a more agile approach has been gathering steam, and businesses that are building agility into their processes are seeing the payoffs. The results are greater productivity, faster time-to-market (TTM) and increased consumer satisfaction.  

    The Culture of Agile 

    Digital transformation succeeds in a business environment that embraces a more collaborative and flexible philosophy. Moving from simply doing agile business processes to being agile across the board requires a shift in the corporate mindset. When a business develops a truly agile architecture, everyone in the enterprise is involved in delivering solutions; from business operations and tech to marketing, finance, compliance and security. The agile culture moves from a company-oriented one to a people or customer-oriented culture, embracing the tenets of collaboration, co-creation, ownership, flexibility and communication.

    Short-Term Cycles and Fluid Planning

    Agility is redefining enterprise architecture from the traditional schematic of centralized governance, lengthy development cycles, scheduled quarterly reviews and reliance on a waterfall delivery method. Instead of long-term operating and strategy plans, agile enterprise architecture involves short-term cycles and fluid planning that is customer and market-centric. The language changes from top-down project mandates and management to continuous product development, experimentation, testing and monitoring within collaborative product teams. 

    Essentials for Crafting Your Agile Enterprise Architecture 

    The blueprint for a truly agile architecture requires fundamental shifts in the business dynamic. There are three essentials that stand out as requirements for attaining agility across the entire enterprise. 

    • Collaboration – The bedrock of successful enterprise-wide agility is collaboration. Innovation will flourish when it is decentralized, and isolated silos give way to cross-functional, agile and self-organizing teams. An isolated IT team leads to delayed projects, overrun budgets, productivity that is hard to measure and disconnects between business and operations. Every department must be involved in supporting and achieving key business goals. 

    Teams containing a mix of business line and IT professionals accelerate development and delivery, greatly reducing time to market. Based upon a shared customer-centric goal and vision, there is shared ownership of outcomes and a deeper level of engagement throughout the enterprise. Daily communication and collaborative feedback nurtures creativity, problem-solving, and drive continuous integration and continuous development. Collaborative input from colleagues, inside and outside the organization, is essential for enterprise, technology, and solution architects. 

    • Current state mapping – Adopting agile practices and principles throughout an enterprise entails potentially massive change. Without a clear roadmap for navigating the process, a business can flounder. Instead of achieving a leaner simplicity, there is increased complexity and a muddied vision. Before a business can successfully embrace an agile architecture, the current state of operations must be fully understood. Essentially, a blueprint of functionalities, capabilities, applications, and the tech stack must be crafted. Everything, from the data catalog to business process workflows must be mapped. Business leaders can ask the following questions to analyze the current state and assess impact points:
    • Which projects should we prioritize and what capabilities will they require?
    • Which processes rely upon what software applications?
    • If we remove one or two apps, how will that impact processes?

    Once the blueprint of the current state is mapped out, a business can work in a structured way to create the future state of its architecture with ease and at its own pace. 

    • Balance agile transformation with daily operations – Continuing to work with some of the current architecture for a time can allow the top level of the organization to ease into adopting the agile mindset and culture. In the iterative process of developing agile architecture, the architects, executives, and tech leaders will need to strike a balance between transforming initiatives and effectively maintaining daily operations. They will have to discern how much direction is needed versus support and how much they will set the pace.

    If the focus is first on value delivered to customers, then the architecture will be prioritized according to the most strategic projects, capabilities and applications. An agile architecture will adopt a ‘just enough’ approach to project design and delivery. 

    Agile Architecture Optimizes Collaboration

    Enterprise agility facilitates more direct communication and involvement of enterprise architects within the cross-functional teams. Through daily discussions and various touchpoints, architects can provide guidance, advice, and feedback on IT decisions and assist teams in jointly owning the solutions. 

    Working together, enterprise and solution architects can develop a complete view of the IT landscape and how it will need to be modified to fit into future needs; minimizing waste and optimizing resources. Solution architects can view the architectural map across all apps, solutions, and systems and collaborate on improvements. They can also help enterprise architects scale standardization across the entire business. That cornerstone of collaboration at the core of agile transformation enables architects and non-tech leaders alike to gain more clarity and alignment.

    The Value of Agile 

    Business value is greatly improved when enterprise architecture combines agile practices. According to a report from McKinsey & Co., businesses transitioning to more agile models can improve financial performance by 20-30%, with 30-50% improving operational performance, TTM, and customer satisfaction scores. Embedding an agile model within processes, people, and technology will cultivate ongoing support for CI/CD, groundbreaking innovation, synergistic collaboration — and ultimately maximize customer outcomes.

  • Languages and DevOps: Recommendations

    Languages and DevOps: Recommendations

    In light of the points brought up in these articles, what is my advice for which language to use on a given project? The same advice I have always given development teams: Use the right tool for the job. My father was a woodworker, and drilled that concept into me. I brought it to development, and it is true. The difference is that “the right tool for the job” is a much more complex question in an existing IT environment than it is in woodworking. Whether you need a chisel or a screwdriver is immediately obvious. Whether Python or Go is the right tool … not so much. Things to consider:

    Does the tool have a broad developer base?

    This is the most straightforward question on this list. Can you get people to work in the language? I worked at a place that still had RPG III code running on a mainframe. We used to joke that those two retired developers in Australia were getting rich from maintaining it – and we still did some new development in the language. It was a bad choice that we moved away from. Maintaining the old while developing a new platform made sense. Packing any new functionality into a tool about which few knew the ins-and-outs, and even fewer were being trained for? Made no sense.

    Does the tool specialize in, or have modules/add-ons that customize it for your target?

    While at the 50,000-foot level, you can theoretically do any computing job in any language, that doesn’t make your preferred language a wise choice. Making certain the tool supports what you need to do is paramount. This is the closest to “right tool” thinking – does it do the job?

    Is the developer base growing or shrinking?

    While this is more important for longer-lived projects, you do want that developer base mentioned in the first point, above, to still be around. This extends to add-ons, also. If you choose ASP.NET today, will it be a solid choice tomorrow? (I’m not saying it won’t, here, just picked one framework to illustrate the point with).

    Does the tool do what you want, or will you be working around faults to get there?

    Honestly, you can – and people have done it – write websites in C. That doesn’t make it the right choice. You’ll end up spending more time getting around the fact that C is ill-suited to this task than you will developing the actual site. Assess the language with regard to how well it actually does the job you need to do. This is a different take than the second point, above – this is the “just because you can, doesn’t mean you should” part.

    Does the tool fit in your environment?

    This is the one that far too many struggle with. New development teams want to bring the tool they’re most comfortable with, because it reduces development cycles in the short term. The issue here is simple. It is far easier to manage an environment that makes a conscious effort to keep our burgeoning number of tools, languages and libraries as controlled as possible. Introducing Go to develop servers, just because someone is excited about it, makes no sense if the entire environment is Node on the server side. Use the environment you have so you can maintain things more easily. Everything is more thoroughly vetted – from which add-ons to include through security – with the language an organization uses the most; take advantage of that. Start with, “We use Java in 70% of applications,” and then deviate from there only if you are moving toward a goal of moving off of Java (again, not saying you should, just an example).

    As a New DevOps Person, Which Should I Learn?

    Not so long ago, the ubiquitous advice to new developers was, “Learn Java!” It was everywhere, doing everything, even when it wasn’t the best tool for the job.

    Similarly, that answer is now “Learn Python!” Of all the languages out there, it is the one used the most. I do not have data to back the following statement up, but I will bet most Dev people have used Python on one or more projects already, and a fair number of Ops people have, too. If Java is the past, Python is the future.

    That said, check out companies where you would really enjoy working and see what they’re doing. Generally speaking, they’re not looking to hire Python developers if they are doing 70% of their coding in Go. I can’t recommend tailoring your learning to one or two organizations’ environment(s) (unless you already work at one of them), but consider the popular languages that have been listed in this series and learn the ones your target organizations use first.

    All of these languages have free/cheap options for learning them. No one is stopping you from learning Go; indeed, there are websites whose purpose is to help you learn by doing. The same type of learning is available for most languages, with implementation details varying a bit. There are a wealth of great books available for all of them, also, if that is your preferred method of learning a new language.

    In the end, learning a new language will never be a negative. It will help you be aware of the list above, and provide meaningful input in the inevitable “Is this the right tool?” conversations.

    So, start with Python and then work through others. With one side note: if you lean toward the Ops side, know what different Ops tools require, and consider learning some of those, also. There is one prominent Ops platform that still uses TCL, for example, and we all know about CICS. So be aware of what might be needed, and at least be familiar with that.

    And Keep Rocking It

    Modern IT environments are unbelievably complex. From rights management to storage, to data gravity, to a wide array of languages and appliances – you and your teammates are keeping it all humming. Keep doing it. Learn new languages before they become necessary to your job, but never think you’re falling behind. If the company is still clipping along, you are the right person for the job.

  • Working with SAFe and Microservices: Learnings from Developers

    Working with SAFe and Microservices: Learnings from Developers

    As developers, the prospect of working for a big company might bring about feelings of oppression, rather than opportunity. Envisaging endless team syncs, pointless meetings and processes wrapped up in bureaucracy inside a set of procedures. It’s easy to imagine why some are put off. For many though, the fear comes from the thought of being lost in a corporate structure where input is not valued, and skills are wasted.

    But every day, as developer concepts become further rooted into corporate behaviors, the line between the two worlds continues to blur. Products are no longer dreamed up in product silos, but from agile teams made up of a diverse cross section of a company. Marketing has dropped drip campaigns in favor of A/B testing, rapid iterations of messaging and data-driven analytics. Big companies no longer mean old school thinking, but fast-changing, inclusive places to really make an impact and be heard. 

    Working within the Scaled Agile Framework (SAFe) with an agile architecture of microservices helps us to do our jobs, but the learnings we’d like to share in this article are universal. An organization may not require the SAFe methodology, but the experiences and methods of working can be applied to many roles and are based on three fundamental questions every developer should ask themselves. 

    How Do I Plan My Day? 

    As with any role, a good start for us is knowing that the previous day went well; that the code we wrote meets the standards required for deployment. Before that, our individual microservices teams will meet to discuss immediate priorities. We let one another know what’s top of mind for us, and any issues we might be facing — and we resolve them together before a midday Scrum meeting.

    The Scrum meeting involves the Scrum master, product owner and the developer team and gives us a chance to give feedback and dive into the wider challenges or blockers we’re facing with our own tasks. We might be thought of as solitary creatures, but we actually tend to solve problems together. So, while we have ownership of our own tasks, problem solving comes from collaboration.  

    Whether working within the agile framework or not, communication is a huge part of our lives. Our advice to developers everywhere is to be vocal about the need to engage within their organizations. And to business IT decision makers, we know that not every firm has the deep pockets to fund a dedicated DevOps team to implement new tools, change culture or break down silos, but there are many communications and open source DevOps tools that boost collaboration and improve workflows — most are well-known to developers. So, give them the chance to explore how they might bring efficiencies to their work. Help them help the business and maybe suggest trialing some new tools. 

    How Do I Upskill on the Job?

    Developers are problem solvers. We love the opportunity to help each other overcome challenges. Peer-to-peer communication doesn’t just allow us to get back on course quicker, it lets different teams share knowledge and pitch in to solve problems. We don’t mind being transparent with our peers when we face an issue – in fact, we quite like it, as the offshoot of working together often leads to new skills and technical knowledge. 

    Learning the interactions between each microservice, for example, helps us map out the architecture and understand how our own work fits into the bigger picture. The more we talk, the more we see where bottlenecks might occur — and the better equipped we are to achieve larger objectives. As we have complete ownership of our services, from initial coding to deployment, it is our duty to ensure our services perform as they should in production. 

    This ownership is something all developers should aim for. Even when working in a more rigid waterfall approach, opportunities to collaborate should be seized and encouraged — even if the only payoff is personal development. Implementing a system of peer code reviews, for example, not only delivers more robust releases, it fosters a collaborative mindset.

    Do I Need to Be in this Meeting?

    Working with SAFe methodology allows the job to be broken down into iterations linked to business objectives. The most significant of these is determined by meetings that occur every four months, where goals are set and roadmaps established to achieve them. Engineering teams then divide up the features that will be passed on to the developer teams, who break them down into two-week sprints. 

    The sprints are what the developer teams work to, under direction from the Scrum masters and product owners. Breaking iterations into smaller sprints ensures we have defined goals, which keep meetings and interactions on track (i.e., daily Scrum meetings and bi-weekly sprint meetings). Without this focus, there’s a risk of derailing the prep required to focus on writing and reviewing code. If we need a discussion about certain features, we make sure to identify exactly who needs to be in that meeting. Above all, we’re all about optimizing our time together. The longer we spend away from our desks, the less time we have to perform our core roles. 

    Whatever your part in the software development lifecycle, never be afraid of asking the question, “Do I need to be here?” Your energies may be best used elsewhere.

    As you can see, these three questions are critical to developers but can be applied to just about any role in just about any organization. Planning, learning and focusing – seems we’re not so different after all.

     

    Tomasz Rumpca

     

    This article was co-authored by Tomasz Rumpca, Java developer and team lead at Finastra. Tomasz works on FusionFabric.cloud, Finastra’s open development platform solution. He is passionate about Java-related technologies and building user-friendly software. Outside of work, he enjoys spending time outdoors with his friends and family.