Databases

Data Wars: Rise of HCL Informix®

Lawrence Fernandes

February 11, 2025

data wars informix hero

“No one’s ever really gone.” — Luke Skywalker

In a galaxy not so far away, where scalability and performance are paramount, one name quietly makes its resurgence — HCL Informix®. Often hailed as a stalwart of traditional relational databases, HCL Informix has been steadily evolving to meet the demands of modern data challenges. With the release of HCL Informix v15, briefly explored on the last Data Wars episode [1], a new chapter begins, one that positions it as a Very Large Database (VLDB) powerhouse, blending its rich RDBMS legacy with other pre-existing innovative features that nod to the NewSQL paradigm. But can HCL Informix truly claim a seat at the NewSQL table? Let’s find out!

A NewSQL Hope

Before diving into HCL Informix, let’s discuss NewSQL. Why does it matter? Or should I say, “Does it still matter”?

First things first, the NewSQL term was first used by 451Group analyst Matthew Aslett in a 2011 research paper [2] discussing the rise of a new generation of database management systems designed to combine the scalability of NoSQL with the ACID guarantees of traditional RDBMS. Back in 2020, I wrote an article [3] about the emergence of NewSQL databases as a solution to the limitations of both traditional RDBMS and NoSQL systems, and concluded by predicting a significant growth in this industry segment.

Regarding my prediction, if we take the financial performance of two of the biggest NewSQL providers as a reference, namely CockroachDB and SingleStore, I’d say I was partially right. According to Sacra, Cockroach Labs’ revenue has grown at a CAGR of 140% from 2020 to 2021 [4], while SingleStore’s ARR grew by 29% from 2022 to 2023, with a valuation of $1.30 billion as of 2023 [5]. Meanwhile, according to Verified Market Reports, the NewSQL database market was valued at $22.81 billion in 2023, and it’s expected to reach $111.14 billion by 2030 with a CAGR of 21.78% [6]. Those are all good numbers, but not exceptional, and far from the absolute dominance expected by many back in 2020.

Now, what about my first claim? Well, while there is a general consensus that NewSQL systems are impressive, aiming to provide horizontal scaling and high availability (the biggest goals of the NoSQL movement) while keeping support for ACID transactions and SQL (some of the best benefits of RDBMS), the reality shows that all that glitters is not gold [7]. NewSQL providers faced many challenges such as market education, integration with existing data ecosystems, compatibility with legacy applications, issues in proving cost-effectiveness in large-scale deployments, failures in guaranteeing consistency [8], and lack of standardization [9][10].

RDBMS (SQL) vs NoSQL vs NewSQL comparison by Dr. Rabi Prasad Padhy

RDBMS (SQL) vs NoSQL vs NewSQL comparison by Dr. Rabi Prasad Padhy [9]

 

Although the NewSQL movement is already mature by now (with over 10 years of existence [11]), and adoption in the market has gained traction, even the top players have shown moderate growth with a limited market share. In fact, most of the original NewSQL providers went out of business, were sold (and failed to land big exits) or pivoted [12][13]. Moreover, the increasing competition from RDBMS providers, and the fact that NoSQL providers fared better in comparison, makes the future of NewSQL uncertain, with many experts already declaring the death of NewSQL as early as 2021 [7][11][12][13].

Competition Strikes Back

Back to 2025, life is far from easy for the remaining NewSQL providers in the market. Selling databases is undeniably challenging — a reality I can attest to from personal experience. The core issue lies in the “stickiness” of databases; enterprises are understandably cautious about migrating from their established RDBMS or NoSQL systems. Furthermore, many of these systems have evolved into multi-model databases, a category that Gartner placed at the Plateau of Productivity in its 2023 Hype Cycle for Data Management.

Gartner Hype Cycle for Data Management
Gartner Hype Cycle for Data Management, 2023

 

This brings us to HCL Informix: a well-established RDBMS with a rich history and pedigree that has evolved into a multi-model database. Founded in 1980 and going public in 1986, HCL Informix (then called just “Informix”) rose to prominence in the 1990s, becoming the second most popular database after Oracle. According to Art Kagel, during the fierce database wars of the ’90s, Informix and Oracle competed intensely for the title of “Best” OLTP performance, with Informix never losing a comparative benchmark against Oracle, Sybase, SQL Server, or other competitors [14]. Speaking of performance benchmarks, an old TPC-D benchmark revealed that Informix was 70% faster than Oracle while running on 25% less hardware [15].

TPC-D Benchmark from the 90s
TPC-D Benchmark from the 90s

 

From 1996 to 2000, the database world legend Michael Stonebraker became Informix’s CTO, after their acquisition of Illustra [16]. Despite its technological advancements, Informix’s success was marred by some technical setbacks [17][18] and financial scandals[19], which led to the downfall of the company. Efforts to recover through restructuring and acquisitions ultimately failed, culminating in Informix’s acquisition by IBM in 2005.

As stated by Art Kagel, “IBM has made more improvements, enhancements, and added more new features to the product than Informix Corp. did during the 18 years of its existence” [20]. Some of those enhancements includes the Informix Warehouse Accelerator (IWA), support for the MongoDB API and connectivity protocols, among
others. However, due to many reasons (some of them explored in the same Quora thread [20]), the competition strikes back, resulting in loss of market share and awareness.

Rise of HCL Informix

Back in 2017, IBM signed a long-term partnership agreement with HCLTech, one of the world’s largest consulting companies (among other related businesses), to jointly develop the Informix family of products [21][22], which bring us to HCL Informix: the same Informix database customers learned to love, but licensed by Actian, a division of HCLSoftware [23].

HCL Informix has product parity with IBM® Informix® Advanced Enterprise Edition, including support for IWA, and now it’s own HCL Informix 4GL and ISQL offerings, as well as exciting new capabilities released in version 15. With a simplified per-core licensing model, competitive pricing, experienced customer support, and no cloud vendor lock-in, there is more to fall in love with HCL Informix [24].

But you may ask, “Ok, but is HCL Informix a NewSQL system?”. The short answer is no. However, HCL Informix has evolved over the decades to include multi-model capabilities (supporting relational, document, time-series, and spatial data), making it able to rival to both NewSQL and NoSQL systems. Its unique combination of traditional relational database features with modern capabilities like high availability, scalability, and hybrid data handling makes it a formidable competitor in both categories, positioning it as a versatile choice for enterprises seeking to bridge the gap between structured and unstructured data management.

HCL Informix’s MongoDB API allows developers to natively leverage MongoDB-like document storage and querying capabilities, also supporting the MongoDB shell and any of the standard MongoDB command utilities and tools, as well as providing a REST API, MQTT connectivity, and JSON data sharding [25]. By providing a unified platform for both SQL and NoSQL workloads, HCL Informix eliminates the need for separate DBMS’es and reduces operational complexity. This hybrid approach is particularly valuable for businesses managing mixed transactional workloads, that also require JSON document storage.

HCL Informix's wire listeners architecture

HCL Informix’s wire listeners architecture

 

HCL Informix’s data replication features allows for seamless replication of data across nodes, ensuring data consistency and fault tolerance [26]. HCL Informix offers a comprehensive suite of data replication features—Enterprise Replication (ER), High-Availability Data Replication (HDR), Remote Standalone Secondary (RSS), and Shared Disk Secondary (SDS)—that rival the replication capabilities of NoSQL and NewSQL systems.

Enterprise Replication enables asynchronous multi-master replication across geographically distributed environments, making it comparable to NoSQL solutions like
MongoDB’s replica sets [27]. HDR, on the other hand, provides synchronous replication for a primary-secondary setup, ensuring strong data consistency, much like NewSQL databases such as SingleStore, which prioritize CA (Consistency and Availability) in the CAP theorem [28].

HCL Informix's Enterprise Replication vs HDR

HCL Informix’s Enterprise Replication vs. HDR

 

RSS adds flexibility by allowing read-only replicas in remote locations, optimizing for disaster recovery and global read scalability, akin to MongoDB’s read preference settings or Spanner’s regional replicas [29]. Finally, SDS extends scalability and fault tolerance by enabling a shared-disk architecture with minimal latency, making it ideal for high-performance OLTP workloads [30].

HCL Informix's HADR vs. ER features

HCL Informix’s HADR vs. ER features

 

HCL Informix’s VLDB capabilities further strengthen its position against both NoSQL and NewSQL systems in handling massive datasets. SmartBLOBs enables the management of large volumes of unstructured data by providing efficient storage, retrieval, and manipulation capabilities for multimedia content, documents, and other BLOB/CLOB data types, seamlessly integrated into transactional workflows—a functionality further enhanced with the release of External SmartBLOBs in HCL Informix 15 [31].

Furthermore, HCL Informix 15 has larger row and page addresses: it’s table size has increased 134 million times, chunk size by 2.25 million times, and storage capacity by 4.2 million times compared to HCL Informix 14.10. This change made the max storage capacity of a single HCL Informix 15 instance goes to half a yottabyte, or 4X the estimated size of the internet!

HCL Informix 15 Re-Architected for Massive Storage Capacity Improvement

HCL Informix 15 Re-Architected for Massive Storage Capacity Improvement

 

Together, these features provide enterprises with a versatile replication toolkit and the ability to handle massive amounts of data, bridging the gap between traditional RDBMS reliability and the horizontal scaling of modern NoSQL and NewSQL platforms. For a deep dive comparison of HCL Informix against some key NewSQL providers, please check the tables below:

Feature/Metric Informix NewSQL Providers
Category Traditional RDBMS with multi-model capabilities Hybrid databases combining RDBMS features with NoSQL scalability
Architecture Traditional RDBMS architecture with optional clustering and enhanced cloud capabilities Distributed, cloud-native architectures built for horizontal scalability
Data Models Supported Relational, Document (JSON), time-series, spatial CockroachDB: Relational

SingleStore: Relational, Key/Value, Document (JSON), Object Oriented, Multi-Value, Vector

Spanner: Relational, Key/Value, Vector

YugabyteDB: Relational, Key/Value

Transactions per Second (TPS) 2 million TPS CockroachDB: 1,684,437 TPS

SingleStore: 10M TPS

Google Cloud Spanner: 1B TPS

YugabyteDB: 100K

Maximum Storage Capacity Half a Yottabyte CockroachDB: ~890TB (10TiB per node, recommended 81 nodes max)

SingleStore: Unlimited (theoretically unlimited on
the Unlimited Storage Tiers, as it offloads data to cloud object storage – probably AWS S3)

Spanner: ~850TB (10TB per node, 85 nodes max – but users can request increase)

YugabyteDB: ~1000TB (10TB per node, 100 nodes max tested – but handle more, but performance bottlenecks may occur)

Tables per Database 477.102.080 (maximum tables per system, supporting up to 21M databases per system) CockroachDB: Virtually unlimited

SingleStore: Virtually unlimited

Spanner: 5000

YugabyteDB: Virtually unlimited

Columns per Table 32K CockroachDB: 1600

SingleStore: 4096

Spanner: 1024

YugabyteDB: 1600 (limit on PSQL, not on YugabyteDB itself)

Scalability Designed for vertical scaling (scale-up) by leveraging stronger hardware; horizontal scaling (scale-out) supported but requires specific configuration

Horizontal scaling features includes fragmentation, sharding (both local or remote shards, and JSON data shards), replication, and distributed queries

CockroachDB: Built for horizontal scaling; auto-shards data across nodes, with effortless addition/removal of nodes

SingleStore: Supports both vertical and horizontal scaling, optimized for fast queries and mixed OLTP/OLAP workloads

Spanner: Exceptional horizontal scalability; scales across regions and zones while maintaining global consistency

YugabyteDB: Highly scalable, horizontally scales by adding nodes, designed for global distribution

ACID Compliance Fully ACID-compliant, ensuring robust transactions in single-node and distributed setups CockroachDB: Fully ACID-compliant, even in distributed transactions

SingleStore: Provides ACID compliance, but transactionality may vary depending on the specific table types

Spanner: Fully ACID-compliant with TrueTime-based strong consistency

YugabyteDB: Fully ACID-compliant, designed for distributed transactional workloads

CAP Theorem Primarily adheres to CA (Consistency & Availability) suitable for OLTP workloads in stable networks CockroachDB: Focuses on CP (Consistency & Partition Tolerance), sacrificing Availability in certain partition scenarios

SingleStore: Aims for CA, optimized for performance over strict partition tolerance

Spanner: Balances CP, with globally consistent transactions using TrueTime

YugabyteDB: Prioritizes CP, offering strong consistency in distributed setups

Ease of Use Mature, enterprise-ready ecosystem with a broad range
of tools and integrations, and strong documentation
CockroachDB: Easy to set up with SQL familiarity
and intuitive tooling; seamless horizontal scaling may require some expertiseSingleStore: Designed for ease of use with integration-focused features (e.g., pipelines), but managing table types (rowstore vs. columnstore) adds complexity

Spanner: Simple for basic operations, but advanced features require understanding of TrueTime and global distribution

YugabyteDB: Familiar SQL syntax and good developer documentation; distributed deployment may be challenging for beginners

Target Use Cases Ideal for OLTP, IoT data, time-series workloads, and traditional business applications requiring stable performance

Support for OLAP workloads thought IWA or real-time integration with Actian Data Platform for Big Data analytics

Support for hybrid architectures: HCL Informix is available on-prem and in cloud marketplaces (AWS, Azure,
HCLSofy). IBM Informix is available on-prem and at IBM Cloud Pak for Data

CockroachDB: Strong fit for global distributed OLTP
workloads, multi-region setups, and modern SaaS applications, and hybrid architectures (on-prem, AWS, Azure, GCP, DigitalOcean)SingleStore: Optimized for mixed OLTP and OLAP use cases, real-time analytics, and fast ingest workloads like IoT. Managed service (SaaS) limited to AWS

Spanner: Enterprise-scale global applications with strict consistency and high availability needs. Managed service (SaaS) limited to GCP

YugabyteDB: Cloud-native, distributed transactional workloads, and hybrid-cloud architectures (on-prem, AWS, Azure, GCP)

 

Thanks for reading, and if you’re interested in HCL Informix, Actian, the Data & Analytics division of HCLSoftware, is ready to support you in your database modernization journey. Until the next Data Wars episode, may the force (of data) be with you! If you liked this blog, consider subscribing to my Data Wars newsletter on LinkedIn.

Honoring a Legacy of Excellence in the Informix Community

I dedicate this article in memory of Harry Carlton Doe III, a pillar of the Informix community whose dedication and expertise inspired countless professionals. Though we never met, Carlton Doe’s contributions—which includes being a founding member of the International Informix User Group (IIUG) and his many Informix books (some of which I own a copy)—have left an indelible mark, and his legacy continues to guide and empower the whole Informix community.

OBS: Informix is a trademark of IBM Corporation in at least one jurisdiction and is used under license.

References:

[1] Lawrence Fernandes. Data Wars: 2024 Wrap-Up.
[2] Aslett, Matthew (April 6, 2011). “What we talk about when we talk about NewSQL”. 451 Group.
[3] Lawrence Fernandes. Data Wars: A NewSQL Hope.
[4] https://sacra.com/c/cockroach-labs/
[5] https://sacra.com/c/singlestore/
[6] https://www.verifiedmarketreports.com/product/newsql-database-market/
[7] https://dev.to/arctype/too-good-to-be-true-why-newsql-failed-l7p
[8] NewSQL database systems are failing to guarantee consistency, and I blame Spanner. Daniel Abadi. DBMS Amusings. September 21, 2018.
[9] NewSQL Databases. Mandeep Kumar. July, 2022.
[10] Google Spanner:A NewSQL Journey or Beginning of the End of the NoSQL Era. Dr. Rabi Prasad Padhy. October, 2018.
[11] Ten years of NewSQL: Back to the future of distributed relational databases. Matt Aslett. June, 2021.
[12] Andrew Pavlo. The official ten-year retrospective of NewSQL databases: Video
[13] Andrew Pavlo. The official ten-year retrospective of NewSQL databases: PDF
[14] Art Kagle. Quora. What are the advantages of using an Informix database instead of an Oracle database?
[15] Informix IDS vs Oracle: A Competitive Comparison. https://slideplayer.com/slide/6229837/
[16] https://en.wikipedia.org/wiki/Michael_Stonebraker
[17] Informix admits faulty code will crash Universal Server. TechMonitor, CBR Staff Writer, October, 1996. https://www.techmonitor.ai/technology/informix_admits_faulty_code_will_crash_universal_server
[18] New Era is not gone, just no longer relevant. TechMonitor, CBR Staff Writer, July 1997. https://www.techmonitor.ai/technology/informixs_new_era_is_not_gone_just_no_longer_relevant_1/
[19] Steve W. Martin. 2005. The Real Story of Informix Software and Phil White: Lessons in Business and Leadership for the Executive Team. Sand Hill Publishing.
[20] Art Kagle. Quora. What are the pros and cons of using Informix as a database? https://qr.ae/pYsl2Z
[21] Art Kagle. Quora. What is the Future of Informix? https://qr.ae/pYsdFQ
[22] https://virtual-dba.com/blog/explaining-the-ibm-hcl-partnership/
[23] https://www.hcl-software.com/actian/informix
[24] https://www.actian.com/databases/hcl-informix/
[25] https://help.hcl-software.com/hclinformix/15.0.0/json/json.html
[26]https://docs.deistercloud.com/content/Databases.30/IBM%20Informix.2/Replication
[27]https://docs.deistercloud.com/content/Databases.30/IBM%20Informix.2/Replication/ER.xml
[28]https://docs.deistercloud.com/content/Databases.30/IBM%20Informix.2/Replication/HDR.xml?embedded=true
[29]https://docs.deistercloud.com/content/Databases.30/IBM%20Informix.2/Replication/ER.xml
[30]https://docs.deistercloud.com/content/Databases.30/IBM%20Informix.2/Replication/RSS.xml
[31] https://help.hcl-software.com/hclinformix/15.0.0/1infocenter/new_features_ce.html#concept_v15.0.0.0__ext_sbspace_15.0.0.0

Disclaimer:
I am not affiliated with, nor endorsed by, any of the authors cited or The Walt Disney Company. References to Star Wars are purely a fan-made tribute.

lawrence fernandes headshot

About Lawrence Fernandes

Lawrence Fernandes is a seasoned data professional with a strong background in data engineering and architecture. With over a decade of experience in the data field, including at global enterprises like IBM and Nestlé, Lawrence has developed expertise in designing scalable data solutions that drive business value. As a Sales Engineer at Actian, he is currently responsible for Latam, working closely with business partners across the region to help customers solve their biggest data challenges. His passion lies in helping companies transition from legacy systems to modern data platforms, ensuring efficiency, scalability, and innovation grounded in sound architectural principles. Lawrence holds a B.Sc. Computer Science degree from CEFET/RJ, and is based out of sunny Rio de Janeiro, Brazil.