As data volumes and value grow, they attract accompanying applications, services, and infrastructure to optimise system performance. Over time, this process can take on its own momentum in a situation analogous to gravitational force, with components drawn together around a central data mass. However, as these systems increase in size and take on greater importance, they almost always become more costly to maintain or relocate.
This is the basis of the ‘data gravity’ concept, which, for many organizations, is having an adverse effect on their cloud strategies and adds more complexity to business continuity provision.
On the face of it, consolidating data, applications, and workflows under a single cloud provider is an attractive, straightforward option. Apply the data gravity arguments, however, and there is often a complex and expensive problem lurking in the background: migrating data to alternative cloud providers or repatriating it to on-premises infrastructure can be extremely costly, slow, and complicated.
Beyond the financial implications of egress fees, concentrating data in one place can also increase exposure to outages, service degradation, or pricing changes, while making it harder for organizations to adopt innovative services from other providers. Deterred by some or all of these issues, some organizations delay or cancel cloud transformation plans, falling into a classic vendor lock-in scenario.
Driven off course
Unfortunately, the consequences of cloud-based data gravity are not confined directly to IT. Agility, innovation, and competitiveness can all be compromised if cloud infrastructure becomes too rigid to support business objectives. For example, if migration plans are slowed or even abandoned due to the cost and complexity of moving data, key initiatives can lose momentum. In some cases, strategic projects are postponed at board level because data gravity has created fundamental barriers to change and, ironically, is acting as a brake on the very innovation that cloud adoption is intended to deliver.
These issues also apply to many organizations on their AI development and integration journeys. The explosive growth in AI workloads and rush to deploy has real potential to exacerbate cloud-based data gravity. On a basic level, AI data needs to be close to AI compute – a requirement that can reinforce lock-in. Training, fine-tuning, and model updates drive continuous demand for high-performance, low-latency datacentre services; and those organizations working with cloud providers on AI projects need to guard against data gravity issues.
The concentration of workloads within a single cloud environment also creates potential vulnerabilities. System outages, cyberattacks, or unexpected price hikes can cause major problems when business-critical data and applications are tied to a single provider. This has led many organizations to explore resilience strategies that build in redundancy, diversify providers, or establish alternative environments in which workloads can be redeployed if circumstances dictate. In this context, the objective is to mitigate the risks associated with over-reliance on the single provider model, which needs to be front of mind for those responsible for business continuity strategies.
Achieving escape velocity
So, for organizations concerned about their exposure to data gravity, what are the options? At the top of the list should be to work out what data exists within the organization, where it resides, how important it is, and how frequently it needs to be accessed. From there, leaders can decide which datasets are suitable for high-performance cloud environments, which can be tiered to lower-cost storage, which may need to remain on-premises, and which can be deleted. By segmenting data in this way, organizations create more flexibility in how it is managed, how backups are organized, and reduce the risk of everything being pulled into a single system.
This might mean using public cloud for scalability and innovation, private cloud for sensitive workloads, and on-premises systems for applications that require low latency or support legacy infrastructure. By matching environments to workload requirements in this way, organizations not only mitigate the effects of data gravity but also create a more resilient and adaptable foundation for future change.
In addition, utilising private networking, cloud exchanges, and open standards can reduce the cost and complexity of moving workloads between environments. Meanwhile, predictable pricing and clear governance frameworks limit the impact of egress fees. These measures also support compliance by ensuring that data can be stored and processed in line with local and regional requirements. Taken together, they provide organizations with greater freedom of choice, making it easier to manage the pull of data gravity without losing the efficiency and performance benefits of cloud adoption.
Building on this foundation, many IT teams are redefining their approach to the cloud. For each organization, a bespoke combination of public cloud, private cloud, and on-premises infrastructure can reintroduce the levels of flexibility and redundancy needed to address these challenges head-on.
The author
Terry Storrar, Managing Director, Leaseweb UK
