On-Prem vs. Cloud Data Analytics

Do you know where to analyze your data?

I’m sure you’ve heard the expression, “Data is the new currency.” It refers to how companies in today’s digital age have greater opportunities to monetize their data. However, that can only happen if businesses can reveal their data’s true value to paying customers. Data has also been referred to as a business’s “greatest asset.” Again, the true worth of a company’s data is determined by the insights obtain that lead to enhancing business operations, increasing customer satisfaction, or inventing, enhancing or customizing products.

This is where data analytics comes in. Data analytics (sometimes called “big data analytics”) quickly identifies trends and patterns in massive amounts of data and, in the case of predictive analytics, calculates the probability of future outcomes based on that information. Statistical algorithms are typically used in traditional data analytics solutions, but increasingly, artificial intelligence (AI) is taking over. According to the McKinsey Global Institute, 69% of the of AI use cases are scenarios where neural network techniques provide higher performance or generate additional insights and applications than established analytical techniques, such as regression and classification.[i]


Of all AI use cases are analytics scenarios.

Where to place your analytics to get the most out of your data

Given the great potential that lies inside businesses’ raw data, most companies want to keep it under their control as much as possible. This means storing it for processing, sharing and analysis in their on-premises data centers (either their own or within a third-party colocation facility). Also, corporate and compliance policies may require that the data never leaves the place it was created. So historically, data analytics has been placed close to where the data resides, on-premises.

However, complex modeling engines for big data analytics workloads require high-performance compute and storage resources to analyze data in real time, Also, proximate access to scalable cloud-based AI systems contribute to streamlining and optimizing data analyses. It is the compute and storage scalability of the cloud that makes it so attractive for real-time big data analytics. In fact, according to Market Research Future, the cloud portion of the global data analytics market is expected to comprise almost a third ($26 billion) of the total global market, an estimated $77.64 billion, by 2023.[ii]

Placing analytic processing capabilities at the edge, proximate to data, solves latency, bandwidth and device complexity constraints.
$26 Billion

Represents the cloud analytics portion, or one third, of the total global data analytics market.

At Equinix, we’ve seen our customers deploy these hybrid multicloud data and analytics environments in a couple of ways:

  1. Keeping their data in edge nodes, such as Equinix Data Hub™, on Platform Equinix®, close to users, databases and other applications, while running all analytics on the cloud. The cloud access is enabled by the Equinix Cloud Exchange Fabric™ (ECX Fabric™), which provides direct and secure software-defined interconnection to multiple clouds using high-speed, low-latency virtual connections.
  2. Running part of their analytics against their data at Equinix, but placing the analytics that requires some “heavy lifting,” and/or AI assistance, in the cloud. Depending on their use case, customers can choose to send only the metadata, a portion of the data or the results of the edge analytics to the cloud, resulting in bandwidth cost savings and an efficient hybrid cloud architecture.

How private interconnection factors into accelerating hybrid data analytics

Placing analytic processing capabilities at the edge, proximate to data, solves latency, bandwidth and device complexity constraints.  You can get multi-destination control of your data analytics workloads using a choice of multiple network providers and cloud analytics platforms on a vendor-neutral interconnection and data center platform, such as Equinix. You also gain the agility and scalability needed for greater growth by leveraging cloud-based analytics and AI platforms that have more flexible compute and storage resources on demand.

By harnessing direct and secure interconnection, versus going over the public internet, you can deploy heterogeneous analytics platforms across clouds and hybrid IT environments and interact with and share data in a secure, seamless, timely manner. The following steps will help you create a high-performance, hybrid multicloud big data analytics infrastructure:

  1. Move part of your analytics processing to the edge where data is gathered from user devices and applications for strategic quality of insight and engagement.
  2. Enhance responsiveness of growing, multicloud-based analytics workloads via secure integration at the edge ꟷ lowering latency, leveraging data gravity with edge caches and improving user engagement response time.
  3. Strategically add capacity and redundancy via distributed scaling of data and workloads to an interconnected mesh of digital edge nodes to improve global performance, adding services and partners as needed.
  4. Guarantee timely, relevant regulatory data collection and brand-enhancing personalization by driving policy enforcement decisions to local regions through edge-based service chaining.
  5. Adapt to changing business needs and models across unpredictable events, new regulations, partners and technologies by dynamically rerouting traffic through a mesh of interconnection hubs, adding new services at local edge nodes in a 24-hour global cycle.

By creating a dynamic data analytics infrastructure at the edge that leverages cloud and AI services, you can:

  • Enhance business strategy, operations and execution in real time, using actionable insights.
  • Leverage global, real-time analytics systems to drive, enact and inform business strategies, linking tactical actions to strategic imperatives.
  • Gain real-time feedback on events, products and services to improve and expand business regionally, nationally and globally.
  • Expand data analytics services and bandwidth without re-architecting every few years.

To learn more about how your business can build a hybrid multicloud data analytics infrastructure at the edge, read the Analytics Blueprint.