All companies are striving to be digital these days. But there remains a significant divide between digital-native companies and existing enterprises engaging in digital transformation programs. One important challenge is the large inventory of systems currently supporting their core business processes.
As enterprises pursue digital transformation, the thought of replacing these systems is daunting. Instead of thinking about IT assets in terms of legacy versus modern, enterprises need to think about their tech infrastructure like a city’s infrastructure. While some old bridges are best torn down and replaced by a new structure (rip and replace), others can be modernized.
It makes little sense for companies to rip out essential systems. There's no need to demolish every back-office system. Legacy systems contain a wealth of historical data that can inform and enrich current digital initiatives. Enterprises instead need to consider augmenting these systems to extend and enhance the existing functionality that is vital for capturing data and business processes.
Meeting New Standards
As enterprises undergo this digital augmentation effort, two important new realities must be adhered to. First, customer expectations have changed. The customer no longer sees the experience model as them coming to the store; enabled by smartphones, the store is now in the customer’s hands. Businesses are expected to have a holistic knowledge of the customer’s context at the moment of the transaction. Without this real-time view, they cannot deliver compelling experiences.
Second, enterprises need to manage the internet’s non-linearity. They need to be able to scale, both in the amount of data they apply to a given transaction and in the number of transactions they can handle in real time at any given moment. This clashes with the performance profiles of many existing systems. In the past, it may have been OK for enterprises to know their global inventory status for that day. Now, it has become mandatory for systems to have a complete real-time status of inventory in a store, neighborhood, city, or worldwide -- all while thousands of transactions are occurring -- and to know when new inventory will be received.
The Path Forward
Enterprises don’t need to make all of their transactions or processes real time at once. They should tackle the most impactful ones first. The first instinct is often to put a cache in front of existing systems. While this partially alleviates the pressure on existing systems that may not scale elastically, it doesn’t provide a path to truly real-time transactional systems. Real-time data platforms allow programming and architecture teams to discover new real-time models for more scalable transaction systems. These new applications can incorporate AI/ML, enabling data-driven recommendation engines and decisioning systems on next steps, and providing new opportunities to better serve customers. Over time, these new applications become the real-time systems of record for inventory, sales, and service schedules and often replace the current back-office versions.
These new systems need to continuously synchronize with existing systems. At present, the most flexible and scalable technology for this is Kafka. There are versions available as cloud services, or Kafka can be implemented on-premises. Another option is JMS, which many existing systems support.
There are evolutionary paths into the digital future, no matter the amount of existing application infrastructure a company depends on. Putting in place an architecture for scalability and throughput as well as integration with existing systems is fundamental. The evolving digital world is real time, it is global -- and it's full of possibilities. All businesses must fit into that world. If not, the opportunity for growth and newfound heights of profitability will vanish and competitors will seize market share. Building a transformational bridge is the best path forward to the digital future.
Lenley Hensarling is the chief strategy officer of Aerospike, a real-time data platform. He has more than 30 years of experience in engineering management, product management, and operational management at startups and large successful software companies. He previously held executive positions at Novell, EnterWorks, JD Edwards, EnterpriseDB, and Oracle. He has extensive experience in delivering value to customers and shareholders in both enterprise applications and infrastructure software.