Virtualization Technology News and Information
GridGain 2021 Predictions: In-Memory Computing as a Foundation for Accelerating Digital Transformation

 vmblog 2021 prediction series 

Industry executives and experts share their predictions for 2021.  Read them in this 13th annual series exclusive.

In-Memory Computing as a Foundation for Accelerating Digital Transformation

By Nikita Ivanov, Co-founder and CTO of GridGain

Over the last decade, many applications have experienced transactional workload increases of 10x to 1,000x, a 50x growth in data, and been under constant pressure to deliver streamlined, responsive customer experiences. Traditional computing infrastructures are often stretched beyond their limits under those demands, incapable of scaling effectively to handle the growth in data and the need for real-time responses.

The global pandemic of 2020 compounded these issues. Digitalization became an imperative, as the majority of interactions moved online and businesses were forced to rapidly change their operations. Some had to scale to support massive influxes of online orders. Others had to pivot sharply to embrace digital transformation in an effort to remain efficient when weakened economies hurt their bottom lines.

A growing number of businesses have turned to in-memory computing as a sound and cost-effective foundation to meet these new needs and deliver the speed, scale, real-time intelligence, and automation that businesses require to compete in today's climate.

Below are my top four in-memory computing-driven trends that I believe will continue to shape digital transformations for the coming year.

1. In 2021, more businesses will need to focus on rapidly accelerating and scaling out applications to meet the challenges of digital transformation

In 2020, the COVID-19 pandemic drove many businesses to dramatically scale out and upgrade infrastructure to maintain high application performance in the face of surges in website visitors, delivery requests, sales transactions, video streaming and more. This was especially prevalent for companies that provide food delivery, ecommerce, logistics, and remote access and collaboration services.

Many of these businesses found that the fastest approach to maintaining or improving performance while simultaneously increasing application throughput was to deploy a distributed in-memory data grid (IMDG) - built using an in-memory computing platform such as Apache® Ignite® - that can be inserted between an existing application and disk-based database without major modifications to either. The IMDG improves performance by caching application data in RAM and applying massively parallel processing (MPP) across a distributed cluster of server nodes. It also provides a simple path to scale out capacity because the distributed architecture allows the compute power and RAM of the cluster to be increased simply by adding new nodes.

In 2021, IMC platforms will become easier to use and the number of knowledgeable IMC practitioners will continue to grow rapidly. This will enable IMC adoption to spread across more industries and to a wider pool of companies. As a result, more businesses will be positioned to take advantage of IMC for rapid application acceleration, not just for response to the demands of COVID, but also to meet new strategic and competitive demands as the pandemic threat abates.

2. Digital integration hubs will increasingly power digital transformations

In 2020, enterprises of all sizes continued to push forward with - and even accelerate - their digital transformations. In 2021, many of these businesses will want to leverage their expanding digital infrastructure to drive real-time business processes powered by 360-degree views of customers and their business. This demand will lead to widespread adoption of real-time digital integration hubs (DIHs), also known as API platforms, smart data hubs, or smart operational datastores.

Powered by an in-memory data grid (IMDG), a DIH creates a high-performance data access layer for aggregating a subset of data from multiple source systems. These source systems may include relational and NoSQL databases, data warehouses, data lakes and streaming data that may reside in public or private clouds, on-premises data centers, mainframes, or SaaS applications. The aggregated data in the DIH can be simultaneously accessed by any number of business applications at in-memory speeds. The IMDG powering the DIH supports a range of APIs, including key-value and SQL, and some even offer ACID transaction support.  A synchronization layer, or change data capture layer, between the data sources and the IMDG ensures that the data in the in-memory cache is constantly updated as changes are made to the underlying datastores. 

With an IMC-powered DIH, businesses can launch a variety of critical real-time initiatives, from predictive analysis and recommendation engines to automated business decision making. Whether in financial services, healthcare, logistics or a range of other industries, these initiatives can instantly react to real-time, cross-organization data views.

3. Persistent memory will make its way into business applications

Persistent memory (PM) solutions such as Intel Optane ensure RAM is not flushed when the servers restart. Instead, data that resides in persistent memory is immediately available for processing at in-memory speeds following a reboot. Some persistent memory solutions also offer the ability to insert more PM RAM per server than traditional DRAM, enabling more memory in a single server, which enables in-memory speeds for much larger datasets on the same server. Combining persistent memory technologies with an in-memory computing platform, such as Apache Ignite, can enable a comprehensive infrastructure for achieving high performance and massive scalability of production applications for both existing and greenfield use cases. This will be particularly useful for high performance, 24x7 applications for industries such as financial services and where high performance, cost-effectiveness, and immediate recovery from a server restart is crucial.

In 2021, we will see a significant acceleration of adoption of persistent memory in business applications. This will take the form of code updates to business applications and the adoption of persistent memory as a standard option with an increasing number of third-party solutions.

4. In-process HTAP will experience growing adoption

HTAP - or hybrid transactional/analytical processing - enables simultaneous transaction and analytics processing on the same dataset. HTAP eliminates time-consuming ETL processes required to move transactional data into an analytical datastore and can power real-time digital business models across a range of verticals, including financial services, e-commerce, healthcare, transportation and many more.

"In-process HTAP," a term created by Gartner, is an HTAP system which can apply real-time machine learning to business processes. In-process HTAP combines HTAP with machine learning (ML) models that are trained continuously using the system data. Powered by an IMDG, such continuous learning environments allow real-time updates to the machine learning model used in the associated business applications. For example, to minimize new loan scams, a bank could continuously update its machine learning model of what indicates a fraud attempt based on the incoming real-time data of new loan applications. It can then seamlessly push the resulting updated model into its production business applications.

The maturing of in-memory computing platforms in 2020, combined with soaring demand to roll out applications that can update their machine learning models in real-time, will drive an increasing rate of adoption of in-process HTAP in 2021.


About the Author

Nikita Ivanov 

Nikita Ivanov, founder and CTO of GridGain Systems, has led GridGain in developing advanced and distributed in-memory data processing technologies. Nikita has more than 20 years of experience in software application development, building HPC and middleware platforms and contributing to the efforts of other startups and notable companies, including Adaptec, Visa and BEA Systems. In 1996, he was a pioneer in using Java technology for server-side middleware development while working for one of Europe's largest system integrators. Nikita is an active member of the Java middleware community and a contributor to the Java specification. He is also a frequent international speaker with more than 50 talks at various developer conferences in the last 5 years.

Published Tuesday, January 19, 2021 7:51 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<January 2021>