3 challenges in transitioning to microservices
Learn about common pitfalls that hinder organizations from adopting microservices architecture
May 28, 2020 | By: Ricardo Madan
Microservices architecture (MSA) is a foundational element to many digital transformation efforts. And to this end, we at TEKsystems look at microservices as a technical expression of Lean and Agile principles that enable organizations to enhance customer experience (CX), move faster, lower technical debt, and squeeze-out waste and redundant processes—ultimately creating a more nimble, flexible DevOps environment.
Although those benefits of agility and efficiency typically feature prominently amongst our customers’ transformation objectives, McKinsey & Company published a 2019 report asserting that 70% of digital transformation initiatives—like moving to microservices—don’t reach their stated goals. Why? At the core, although most of these projects have the potential to yield priceless advantages on the back of better responses to the changing needs of customers—the appropriate internal conditioning needed for such a seismic change is often lacking and can doom these initiatives from the onset.
So, when transitioning from a monolithic or legacy system in favor of a microservices architecture, organizations should be mindful of common challenges that may impact their ability to reach desired velocity and ultimately reap the long-term benefits MSA can offer. Like in our Version Next, Now issue on Cloud Enablement, where we focus on improving the value stream by transitioning from monoliths to microservices—we’ll take that conversation further and explore three common challenges organizations may experience when trying to implement microservices.
Like any sort of modernization effort, adopting or implementing microservices doesn’t start with technology. In healthy organizations, it starts with strong IT and business leaders who clearly understand and communicate how the targeted technology is going to drive quantifiable business value and improve life for their customers and internal teams as a result. Andy Jassy, CEO of Amazon Web Services (AWS), captured this sentiment during his December 2019 AWS re:Invent keynote, where he talked about senior leaders in a given organization having to be “convicted and aligned” in order for any transformation effort to succeed, since the absence of such voids the mechanism to catalyze change.
Unsurprisingly, the failed transformations we’ve seen around MSAs or any related modernization can be traced back in almost every instance to leaders who either did not possess or communicate a clear vision, or otherwise failed to set and drive aggressive top-down goals and expectations—all of which, when done, keep the customer at the center of the transformation effort in question. Bottom line: without strong, clear and decisive leadership, modernization efforts are doomed. And, without appropriate leadership resolve, the next subsequent challenges that we cover (organizational and talent) simply can't be dignified.
2. Organizational or political
Flatter, leaner DevOps teams that are built for speed, accountability and experimentation make up the optimal environment for MSAs to succeed. In more enterprises than we'd like to admit, DevOps serves as an empty buzzword versus standing as a pillar of organizational IT culture. If we just look at code and the way code has to flow for a given application in a traditionally siloed environment, there's usually an unnecessarily complex transit of authorship and custody across a given dev→test→deploy life cycle, with each one of those functions often becoming a monolith unto themselves.
Years of Waterfall culture is usually the invisible culprit to this phenomenon, which typically inflates the number of hands in the process and bloats the amount of control exerted therein. We as an industry consequently still deal with the associated hangovers of stifling bureaucracy, territorialism and prohibitive constraints up and down the application/system life cycle. And unless strong leaders see this and are able to cut through resulting mire, their adoption of any meaningful scope of microservices will stall out.
Now, I'm not implying that governance is not essential, but many organizations confuse overweight checks and balances with control and assurance. The incremental insidiousness therein is that over time, many redundant or unnecessary "checks" also become monolithic, and this consequence is eventually on display via massive org charts, overengineered processes and oppressive tollgates, which all have to be satisfied before a single line of code ever sees the light of day in production. The net result is that the entire organization moves slowly in serving its customers. The irony here is that MSA style development delivers faster iterations, faster response to customer feedback and deployment of new/enhanced features, but the leap to embracing and institutionalizing this approach is often arrested due to historic proliferation of organizational or political constructs that stand in the way.
In TEKsystems’ 2020 State of Digital Transformation survey, 39% of respondents say the biggest challenge to digital transformation was a complexity of current environment / siloed mindsets and behaviors.
Don't underestimate organizational change efforts toward adopting MSA. Instead, cultivate a culture that embraces change, agility and speed.
3. Talent and expertise
Critical to the success of any digital transformation or microservices initiative is having the right expertise and talent to execute your roadmap. However, there aren’t a ton of people who possess in-depth knowledge of MSA. Our digital transformation report reveals two in five companies don’t believe they have the expertise to succeed with their digital transformation initiatives. This is a problem—and although a lot of fuss is made about how the lack of IT talent threatens any transformation or modernization effort, experience tells us that organizations can’t simply recruit their way out of the problem.
Especially considering the organizational or political constraints described above along with related technical complexity, organizations who thrive with broad-scale MSA adoption typically have a multipronged strategy that closes the talent and expertise gap. Prevalent across those organizations, you will find a combination of the following:
- Boot camps: Organizations who find a way to harness the power, creativity and perspective of new entrants to IT are cultivating a digitally native workforce from the onset of tech careers. Although the skills, tools and functions across the digital landscape are indeed diverse and can be lent to many different boot camp formats aimed at accelerating the expertise you need to thrive in the future, we’ve found that the genesis to building “native MSA talent” can start with a combination of full-stack web/ops/system development. And although I look forward to more conversations regarding how leading organizations actually solve the long-term talent gap by broadening their STEM candidate pool, check out the following as a primer for full-stack development boot camps: Full-stack Bootcamping.
- Upskilling: Upskilling is nothing new in IT. There’s always a new technology tool, process or architecture that replaces older or outmoded versions, and the best organizations have found ways to institutionalize the associated reskilling or upskilling of their workforce in order to best adapt to the latest and greatest practices. Relating to microservices, the paradigm shift driving this latest and greatest development style is however profound enough that we can compare it to the legions of mainframe programmers who began transitioning to client–server, web and distributed development styles, post Y2K. That being said, we’ve seen the most impactful MSA upskilling programs thoughtfully coincide with target modernization objectives. For example, the associated dev and support teams who “keep the lights on” with legacy applications (often of the monolithic variety) are systematically enrolled and reskilled ahead of interweaving project plans, which ultimately transform their current environment to the MSA target state. That sounds easy, but the inputs can become complex since your organization is essentially embarking on a parallel path beyond the technical modernization at hand, so proper reimagining of your workforce is required in advance. For more info, feel free to check out upskilling with TEKsystems Learning Solutions.
- Transformation partners: Even though microservices eventually get you to a much simpler dev and CI/CD model aimed at propelling CX in a highly efficient manner, initial transformation is complex. Especially considering that MSA typically involves broader integration of container-based deployment, cloud-native services and automated code pipelines—not to mention the Agile transformation required of associated people and process patterns—the best organizations align themselves with consulting and system integration partners who possess the expertise and track record for helping accelerate related transformation objectives in a truly lean and value-driven fashion.
During the same AWS re:Invent keynote address mentioned earlier, Jassy spent valuable time highlighting the critical role that relevant consulting and system integrator partners serve toward all such modernization efforts. And considering the relentless pace of change in and around these new computing spheres, Jassy went on to describe the very best partners as the “new guard” of available systems integrator and consulting companies. Since effective MSA does not require high degrees of overhead or legions of redundant functions performed in off-shore locations, one can argue that the value proposition of the new guard of transformation partner is very different than the legacy consulting and “big box” integrators who’ve dominated the industry’s Waterfall landscapes to date. Whereas the old value props revolved around massive benches of off-shore talent and uber-comprehensive project management models (all for which customers have paid historic premiums, with oftentimes dubious ROI), the new guard of partner is known by their ability to rapidly yet accountably move their customers down the transformation journey. And although that’s easy to say, I would encourage organizations to separate the wheat from the chaff by scrutinizing a potential partner’s past performances—what do their customers say about them? Another good point of scrutiny is a potential partner’s IP—what provable frameworks will they lean on to accountably help you modernize? And of increasing importance are the accelerators, products and codebases that partners bring to bear in order to help reduce cost, time and effort during these journeys. When considering partners, if candidates can’t answer these questions in an audible-ready fashion, much less correspondingly demo related IP and accelerators, keep looking!.
Ricardo Madan is an executive leader of TEKsystems Global Services with over 20 years’ experience developing people, serving customers, designing and building technology services/products, and driving a seismic array of growth venues all with the goal of significantly improving the lives and circumstances of his team, customers and community.