Cloud Migration Woes Need a Data Rethink

LG CNS was facing a gargantuan task.

The IT services provider needed to migrate 70 applications from their on-premise data centers to an AWS Landing Zone in 70 days. Dubbed the “70 in 70” project, their clients depended on its success.

It is a similar scenario played across the world today. Companies, both enterprise users and service providers like LG CNS, are waking up to a clamor for public cloud migration. But with core data and applications running on legacy and on-premise infrastructure, it can be a daunting feat.

COVID-19 has ramped up the volume on cloud migration urgency. Companies need the business agility and scalability that public cloud promises as they walk on operational egg shells. But whether they are lifting and shifting, re-hosting, refactoring, or rebuilding your infrastructure in the cloud, they need to first understand that the cloud is different.

Prepare to change how you operate

The biggest change is the operating model. You need a different one for cloud.

“One of the most critical mistakes we see associated with migrations is the lack of an appropriate cloud operating model that enables the organization to scale its cloud outcomes,” said Andrew Venables, director of strategic alliances for APAC at Contino.

Using the on-premises operating model is ill-advised. “Using existing on-premises operating models will constrain the potential benefits of cloud and provide an inefficient user and technical experience that could disillusion staff,” he added.  

Venables advised companies to shift their approach from project-focused to a product-focused one when migrating. “So that organizations can focus on an end-to-end value stream for better ROI and also cater for modern ways of working.”

Other common issues include trying to plan everything upfront and trying to do the entire migration “all at once.” This will result “in an ‘analysis paralysis’ rather than executing in small iterations,” he said.

The relationship between cloud service providers and customers also changes. In LG CNS’s case, the strong teamwork between LG, AWS and Contino allowed the company to meet the migration goals.

“The close collaboration between the organizations also led to the development of the technical foundation of LG CNS’ cloud consulting businesses, where local consultants and engineers were trained in cloud architecture throughout the project,” said Venables.

Look out for the data gravity

One concept that is now seeing a resurgence is data gravity — a term made famous by software engineer Dave McCrory in 2010. He likened data to a planet. If there is a cluster or concentration of these in one area, like gravity there will be an increase in the attraction of objects surrounding it.

In IT terms, those objects are applications and services. McCrory, who recently joined Digital Realty as its vice president of growth and head of insights and analytics, claimed that while the other “objects” also have their gravity, data has the greatest pull.

COVID-19 made data gravity a huge issue. As companies shifted online and created more data, new gravity “wells” formed.

“COVID-19 has a play in it, but so does digital transformation and organizations looking for new channels of revenues. Organizations are also sensing and collating data [using IoT] and this is creating a new trend [with data gravity],” said Joseph Badaoui, senior engineer for APAC at Digital Realty.

Data gravity shows we need better data distribution, which goes opposite to on-premises practice of centralizing data in a few data centers. A concentration of data in the cloud can impact latency and increase complexity (imagine moving a huge data store across two cloud data centers). Part of McCrory’s claim is that data travels faster nearer to a data gravity source, so it makes sense to keep the right data near to its destination.

This is where a good cloud infrastructure partner helps. Badaoui noted that his company has solutions to ensure that the gravity wells are spread across the infrastructure globally. Hyper scalers like AWS and Google also offer similar solutions and help to distribute data gravity across their data centers. 

No matter which approach a company takes, it is important to plan for proper data distribution. “You must plan. That is how you can optimize for your business needs. Have a strategic play and partner with a strategic player who has done this — one who has the technical know-how and the infrastructure to help with the infrastructure and spread the data gravity,” said Badaoui.

Going into cloud is not an IT matter

Cloud migration can be an immense technical challenge. Yet, it is not an IT matter. Moving an entire on-premises infrastructure sitting inside a company’s perimeter to the public cloud is no trivial matter. The reason is that you are changing the company’s value proposition.

It is why Ajit Melarkode, vice president for Asia Pacific & Japan (APJ) at Boomi believed cloud migration should be a Board-level or senior management concern. 

Moving to the cloud is also not going to solve any legacy infrastructure issues. Melarkode argued that companies need to take a step back to ensure legacy-to-legacy and legacy-to-cloud issues are addressed upfront. 

“The leaders among companies spend more time changing the business and preparing for the future,” he observed, adding that cloud migration essentially involves organization-wide change management. 

It is why Singapore-based smart real estate player Ascendas-Singbridge (ASB), Asia’s leading sustainable urban development and business space solutions provider, used Boomi integration platform to connect and centralize its largest ever operational transformation.

“We knew we needed a digitally-agile environment and exceptional, meaningful data to better understand our clients so we could offer tailored experiences. This is especially important for clients we are helping expand into the Asia-Pacific market,” said Alvin Tan, head of group information management and technology at ASB in a press release.

Outdated connectors have no place in cloud migration, while making it difficult to have a single source of data — a key ask for cloud-based data management.

For example, Scoot, the low-cost arm of Singapore Airlines, used the low-code Boomi integration platform to replace a series of outdated connectors, which did not provide the level of automated data management the organization required.

Its former integrations limited communication between Scoot’s systems and restricted access to data, inhibiting the potential of its sharing capabilities. These bespoke integrations were also code-heavy, consequently demanding substantial maintenance.

“With the various features the platform provides, we will be able to connect our entire organization to create a single source for our data, with the knowledge that this information is up-to-date and accurate," said Jason Chin, Scoot’s vice president of information technology.

"We will then be able to better understand our business and customers, and deliver the products and services that passengers want — before, during and after their flights.”

Whatever the approach, Melarkode noted that cloud migration should be viewed differently and not like another IT problem. After all, “cloud is an opportunity to do things in a new way, and not a destination,” he added.

Photo credit: iStockphoto/HbrH