×

Five Exciting Big Data Trends Worth Taking a Closer Look

By: Kamil Bajda-Pawlikowski
September 20, 2022
Share: Linked In

After Covid-19, many business executives faced one of the toughest leadership tests to turn this challenge into an amazing opportunity. What did the business world do? They decided to turbo-charge their digital transformation plans. Organizations worldwide began identifying opportunities by extracting insights through data lake analytics. This enabled them to discover trends and patterns never before discernable.

With this in mind, here are the five most exciting big data trends we see today:

#1 The adoption of data lakes accelerates

Data lakes have become a highly economical option for companies. The rise in remote and hybrid working environments has increased the need for data lakes for faster and more efficient data manipulation. With Microsoft, Google, Amazon and other tech giants actively encouraging the move to the cloud, the adoption of data lakes is making it easier and cheaper.

As organizations migrate to the cloud and focus on cloud data lakes, they will also move to converge the data warehouse with the data lake.  Data warehouses were created to be optimized for SQL analytics, but the need for an open, straightforward and secure platform that can support the rapid rise in new types of analytic requirements and machine learning will ultimately see data lakes become the primary storage for data.

The adoption of Data Lakes will continue into 2022 and beyond with the market expected to grow from $3.74 billion in 2020 and is expected to reach $17.60 billion by 2026, at a CAGR of 29.9% over the forecast period 2021 – 2026.

#2 Streaming data and data at rest unifies for improved predictive capabilities

Big Data analytics today focuses on two primary sources — streaming data and data residing in a database or data lake. We expect to see these sources continue to converge with streaming and operational systems, providing more unified analytics. The result will be an improvement of data-driven insights to improve operational decision-making through the use of lightweight analytics and improved predictive capabilities.

With a data lake or even a simple database, queries can be fairly complex without regard for dynamic data flows that require extensive resources to process. Streaming data is fluid, and those resource demands and ongoing additions therefore require that its queries remain superficial. As such, today’s predictions for financial markets, supply chain, customer profiling, and maintenance-repair-overhaul – are limited, often based on lightweight, “shallow” data.

We anticipate the steady increase of cloud-based storage and applications providing the elasticity needed to eliminate resource limitations and replace the traditional approach of the familiar centralized structure.

Performing analytics on distributed clusters — and aggregating the results of both streaming and operational data sources on other clusters into a single pane of glass — will become the norm. The results will yield truly comprehensive predictive models, taking the best from a data lake’s deep data and the streaming source’s live data flows.

#3 Data sharing made easier

Beyond the technical advantages of cloud migration (hardware support, storage/bandwidth limits, backup, and security), perhaps the most obvious is the ability to share data that is no longer stored physically within a company’s internal network.

Providing valuable data – of use strategically, financially, or even for compliance – to third parties simplifies and streamlines distribution processes for both the provider and consumer. One significant benefit: the data lake/streaming data analysis discussed above now has a new consumer base. Whether focusing first on a commercialized, public-facing marketplace like AWS, or starting with an internal sharing platform like Snowflake’s (for internal departments and some verticals), this paradigm applies to each approach, and offers fundamental improvements to the complex, multi-step systems and policies in place today.

Cloud providers will offer both these data exchange offerings in order to capture the market for both “intranet and internet” data providers and their consumers.

#4 Smart query engines seamlessly adapts to process unprepared data

Database optimization is being sped up and improved by baking machine learning (ML) right into the database. It’s a prime use case, as the ML has access to its most valuable resource for building effective models: massive amounts of anonymized data, within a well-defined structure and context.

We have witnessed this trend making strides with the creation or dropping of indexes as the query engine senses the need, but this is just the beginning, and momentum will snowball. This trend is seeing an increasing drive towards the separation of data storage and data consumption. The next generation engine will embrace this separation between data storage and consumption by applying dynamically acceleration strategies, such as cache and index, based on the analytical workload pattern and behavior.

The philosophy behind this revolution is to ‘let the engine work for you’. The engine should not expect the data to be prepared, rather the engine adjusts itself to the data it encounters. This wide-open space will become a must-have rather than a nice-to-have function, as customers discover both cost savings and improved performance.

#5 Predictive analytics will drive next-generation of digital applications

As we began this blog with our discussion of the merging of analytics drawn from dynamic data feeds and data lakes, the access to these insights will need to be re-imagined. The classic dashboards used for “data storytelling” are today based on historical data carefully collected, queried, and gathered into a report for periodic review. It’s good stuff, but it’s outdated by the time it’s compiled and presented.

Yes, the dashboard will remain in use, but the content offered will be live and as-it-happens dynamic, drawn from processes built right into application code. Significantly, access to this information will also be democratized across all relevant internal departments, available directly to tactical teams like sales, marketing, QA, and others – rather than having to be parsed, interpreted, and distributed by a data department. With live trend analysis, these departments can adapt and improve much more quickly than with today’s longer-term cycles. With the recognition that business value is often about how people react and behave, rather than simply following the money, this game-changing drive toward prediction is an exciting “perfect storm” of new advances in cloud, database, and analytics.

Looking ahead

The merging of several technological paradigms maturing steadily in the past few years is set to create a less compartmentalized, historical, and resource-constrained analytics ecosystem. Companies with the most to gain are those who value their ability to quickly adjust their processes and services based on what customers tell them they prefer — explicitly, but more and more, passively, through their actions.

Kamil Bajda-Pawlikowski

Co-Founder and CTO, Starburst

Kamil is a Co-Founder and CTO of Starburst. Previously, Kamil was the chief architect at the Teradata Center for Hadoop in Boston, focusing on the open source SQL engine Presto, and the cofounder and chief software architect of Hadapt, the first SQL-on-Hadoop company (acquired by Teradata). Kamil began his journey with Hadoop and modern MPP SQL architectures about 10 years ago during a doctoral program at Yale University, where he co-invented HadoopDB, the original foundation of Hadapt’s technology. He holds an MS in computer science from Wroclaw University of Technology and both an MS and an MPhil in computer science from Yale University. Kamil is co-author of several US patents and a recipient of 2019 VLDB Test of Time Award.

Start Free with
Starburst Galaxy

Up to $500 in usage credits included

  • Query your data lake fast with Starburst's best-in-class MPP SQL query engine
  • Get up and running in less than 5 minutes
  • Easily deploy clusters in AWS, Azure and Google Cloud
For more deployment options:
Download Starburst Enterprise

Please fill in all required fields and ensure you are using a valid email address.