GridGain is a VC-backed software company that was initially formed in 2007, but not incorporated until 2011. In 2014, it donated its code base to the Apache Software Foundation as the open-source Apache Ignite project, and the company continues to provide support for Ignite in addition to its own products.
Headquartered in the Silicon Valley, California, GridGain has additional offices in Canada and Portugal as well as workplaces in the UK, France, Cyprus, and Armenia.
The company has over 500 deployments at large enterprises across various industries, including household names like JPMorgan, AT&T, and American Airlines, and boasts an extensive partner program that spans systems integrators, cloud platforms, independent software vendors, and technology providers.
Company Info
Headquarters: 1065 East Hillsdale Blvd. Suite 220, Foster City, CA 94404, USA Telephone: +1 650 241 2281
Fig 1 - GridGain Unified Real-time Data Platform at-a-glance
The GridGain Unified Real-Time Data Platform combines distributed processing and memory-centric storage to deliver low-latency analytics at scale. The commercial offering builds on the core Apache Ignite functionality, and is available in Community, Enterprise, and Ultimate Editions that provide successive additional features on top of Ignite. It is available for deployment on-premises, on any public cloud (including hybrid-, multi- and inter-cloud deployments) and through GridGain Nebula, either as a self-service SaaS on public cloud or as a fully managed service anywhere. An overview of the product is shown in Figure 1.
Gridgain’s core selling points are twofold. First, it provides a wide variety of unified data functionality. This includes handling data at rest and in motion, on-disk and in-memory storage, transactional and analytical processing, support for streaming and historical data, AI and MLOps, integration with (and thus access to) your system(s) of record, and more. This can dramatically simplify your data architecture, since it allows you to replace many point solutions with one unified platform, reducing the time, effort and money spent on maintaining both a laundry list of data products and the connections between them. Moreover, it will frequently be good for performance, as the majority of your data and data processing will happen over one tightly-integrated platform rather than a much looser collection of solutions that will often require data to be moved between them in order to process it.
This feeds into the second major draw for the GridGain platform, which is that it allows you to operate on your data with very low latency, even for the numerous (and often mixed) types of workloads we have just described. It primarily achieves this via in-memory computing, supported by the platform’s in-memory data grid. In fact, GridGain started life as an in-memory data grid, and as such these capabilities are some of its most highly developed.
This is particularly pertinent for addressing use cases that require real-time, large-scale data processing and decisioning, such as real-time AI, fraud detection, risk management, 360-degree view, and operational analytics. At the same time, in-memory compute alone will not always be enough to operate in real-time while addressing these use cases, as you will also need to contend with latency issues caused by data movement, the need to aggregate data that is held in separate silos, massive scale requirements, and highly complex workloads. As such, GridGain has been designed to address these problems.
For instance, as a unified platform it inherently helps eliminate silos and reduce the need for data movement. Its use of collocated analytics, performed wherever the relevant data resides, reduces the latter even further. We have already mentioned the platform’s ability to operate on a variety of workloads, and it is also highly scalable without compromising its performance.
In short, GridGain has developed into a unified, real-time platform for operating on data at scale and at speed that provides a wide array of data-oriented capabilities, including storage, processing, and analysis, in order to address the business needs surrounding data in a multi-dimensional fashion.
Customer Quotes
“By using Gridgain, we have been able to reduce our pricing strategy computations from many seconds to milliseconds. We have scaled out our computing platform across all geographies to centrally manage our pricing data and provide real-time promotions to our customers.” EssilorLuxottica
“GridGain met our criteria for ACID compliance, support for high volume transactions, horizontal scalability, and co-located data processing.” American Airlines
Fig 2 - GridGain Unified Real-time Data Platform architecture
GridGain leverages a distributed, in-memory data store that eliminates latency caused by moving data to and from disk while at the same time offering serious potential for horizontal scalability. At the same time, it is not exclusively an in-memory platform, and can be deployed as a hybrid-memory database that utilises both in-memory and disk storage in whichever ratio you prefer. Its architecture is shown in Figure 2. The platform also includes immediate data availability on restart (albeit with some initial latency while it is loaded into memory) and the option to use column storage (currently in beta).
It can handle both operational analytics and HTAP (Hybrid Transactional/Analytical Processing) workloads, provided in real-time thanks to (among other things) the in-memory capabilities and collocated analytics described above. HTAP is available both at point-of-decision (with transactional and analytical processing sharing the same data infrastructure) and in-process. Stream processing and streaming analytics are also available, using event-driven data ingestion and processing. To facilitate this, the product provides native integration with Apache Spark, including support for RDDs (Resilient Distributed Datasets) and DataFrames. It also offers a connector to Apache Kafka certified by Confluent, Kafka’s primary developer and supporter.
GridGain provides various means of accessing your data, including SQL queries, key-value APIs, and continuous queries. You can also use it to read and/or write to any integrated systems of record. Transactions are fully distributed and ACID compliant, and options for both immediate and eventual data consistency are provided. It offers automatic integration with its in-memory data grid, including no-code options for configuring schema import, data loading, and connectivity. Its data grid uses SQL to integrate with RDBMSs, and it comes with native support for Cassandra, MongoDB and Hadoop.
There are several reasons to deploy GridGain as part of your solution. In fact, we have already mentioned more than a few, such as the way it can be used to simplify your data architecture and its mitigation of data movement. Highlights include its performance and scalability, particularly the maturity of its in-memory capabilities; its flexibility in interacting with your data, especially its HTAP support and provisions for data access; and its ability to run as a system of record in and of itself, thanks to ACID compliance, high availability features, and strong data consistency.
The Bottom Line
The GridGain Unified Real-Time Data Platform is highly performant and can address a variety of data processing use cases. This should make it appealing if your goal is to streamline, accelerate, and/or consolidate your data ecosystem.
We use third-party cookies, including Google Analytics, to ensure that we give you the best possible experience on our website.I AcceptNo, thanksRead our Privacy Policy