Many enterprises have managed to adopt a digital veneer by tinkering at the edges. But at some point, hard decisions have to be taken about modernizing legacy systems. Here’s how to minimise the risks

The start of a new year is always a time of reckoning, and enterprises that haven’t yet modernised will be wondering whether 2020 is the year to bite the bullet.

For these enterprises, it is likely that for some time now, their existing systems have not performed as their digital-savvy customers would expect, modern technologies have been too difficult to incorporate, and throwing more hardware power at the problem has proved extremely costly.

But the decision to replace, rebuild or re-architect legacy enterprise infrastructure is never an easy one. CIOs know that the task carries considerable risk of service disruption.  So how can the risks be lowered? The advice from the experts is now fairly universal – focus on the required technology outcomes rather than the specific tool upgrades.

This is easier said than done when there are IT giants pushing their respective agendas – cloud, big data, IoT, AI, etc, etc.  Below are some considerations that should be ticked off ahead of any platform modernisation exercise.

1.   Determine your desired business outcomes

It is obvious but worth reiterating that desired business outcomes should drive IT capabilities and not the other way round. Similarly, the length and breadth of an IT modernisation exercise should be driven by an enterprise’s digitization business roadmap.

The roadmap should specifically address those tech capabilities that are already in place and to be maintained, versus new ones yet to be implemented in order to achieve pre-determined business targets. This then enables IT architects to dimension what and when to migrate from one platform to another.

2.    Secure your business logic

The wisdom adopted by the experts is that Big Bang approaches are inadvisable, except in very specific and controlled circumstances. For all others, a multi-phased approach helps minimise downtime, lower the risk of large scale failure and improve end-to-send system testing.

However, it is crucial that the enterprise’s services, and therefore its business logic, remains accessible throughout the extended upgrade period. This requires the logic embedded within legacy mainframes to be exposed. Today, there are patterns for ensuring that new systems can be introduced alongside legacy ones, by splitting monolithic applications into microservices. These serve to ensure that the existing business logic is preserved and legacy systems are not destabilised during migration.

3.   Leverage APIs

Enterprises should have as their mantra that under no circumstances should business outcomes be disrupted, even during legacy transformations.  If so, then a central component of any such transformation exercise is the encapsulation of  all existing business data and functions by leveraging API and microservices technology.

APIs allow enterprises to take a incremental approach to modernisation. Coupled with a router-type mechanism, enterprises can choose to build new microservices on new platforms, move existing microservices to new platforms, and continue to access existing microservices on legacy platforms, all at the same time.

4.   Ensure business flexibility using event architecture 

While ensuring that existing processes are maintained, enterprises also need to transition smoothly to new business processes where required.  Such transitions require, in effect, the ability to integrate the new and the old, whether across technologies, applications or data sources. To this end, event-based integration architectures offer more flexibility than other options.

An event-based architecture is characterised by its ability to capture business events, for example, a new customer registration or a payment transaction. These events are captured as they happen in (near) real time, regardless of source or system. Event-based integrations means that newly-developed services built on next-gen platforms can be easily merged with existing services built on legacy platforms. This ensures business process changes can be made seamless.

5.   Introduce AI gradually

According to Forbes, “Companies that still aren’t investing heavily in analytics by 2020 probably won’t be in business in 2021”. AI and machine learning can certainly be a “force multiplier”, given its potential to engage customers in unprecedented ways.  But many AI adopters have expressed disappointment with the results of their early experiments.

Experts say enterprises should not, in the first instance, use AI to devise new market strategies or business targets. Rather, AI tools should be introduced simply to overcome process road blocks. Only when executives become familiar with such tools should they be applied for more disruptive purposes. This step-up approach to AI also defines the way AI-friendly infrastructure upgrades should take place.

6.   View change as constant 

New technology is emerging at such a breakneck speed that it is impossible today to imagine large modernisation efforts happening only once every few years, as in the past. While there remains a desire to look ahead and thereby “future-proof” current investments, the reality is that few can foresee technology developments even a year into the future, let alone three or five years.

The arrival of new technologies should therefore not be the primary catalyst for making IT investments. In contrast, business imperatives are likely to evolve more slowly and make for a more stable beacon by which to guide legacy transformations. The decision-making process for migrating to new platforms must similarly be based on a combination of business and IT teams, while the implementation process will likely require a vibrant partner ecosystem.

Percipient’s UniConnect 3.0 platform offers an out-of-the-box solution for enterprises looking to migrate from legacy to modern infrastructure. We help build an migration path that accelerates the extraction of new digital and AI-driven business value, while ensuring existing services are not impacted.