MA R K E T

MAP
OCT 2019 Data Platforms Market
Map 2019
James Curtis, Senior Analyst, Data, AI & Analytics

Csilla Zsigri, Senior Analyst - Blockchain & Distributed Databases

As enterprises large and small look to data and analytics as drivers for digital
transformation, the potential value of the category has never been greater. The
2019 Data Platforms and Analytics Market Map illustrates this evolving landscape.

©2019 451 Research, LLC | W W W.4 51RE SE A RCH.CO M

 About the Author
James Curtis
Senior Analyst, Data, AI & Analytics
James Curtis is a Senior Analyst for the Data, AI & Analytics Channel at 451
Research. He has had experience covering the BI reporting and analytics sector
and currently covers Hadoop, NoSQL and related analytic and operational
database technologies.

Csilla Zsigri
Senior Analyst - Blockchain & Distributed Databases

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
II
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
Key Findings
• The combined data platforms and analytics markets generated revenue of $62.4bn in 2018
(excluding blockchain and distributed ledger technology), according to 451 Research’s
September 2019 Data, AI & Analytics Market Monitor, with revenue expected to exhibit a
combined annual growth rate (CAGR) of 7.2% between 2018 and 2023.

• The top 10 data platforms and analytics vendors generated 87% of all the field’s revenue in
2018, according to 451 Research’s Data, AI & Analytics Market Monitor. That figure is expected
to decline to 83% in 2022.

• The fastest rate of growth among the various data platforms sectors between now and 2022
consists of distributed data processing (24%), distributed data grid/cache vendors (a 23.4%
CAGR), non-relational databases (22.8%), event/stream processing (17.1%), analytic databases
(7.0%) and relational operational databases (6.1%).

• According to 451 Research’s Voice of the Enterprise: Digital Pulse, Budgets and Outlook 2019
survey, worldwide, 5% of enterprises are using blockchain in production, 16% are working
on proofs of concept and pilot projects, another 7% have plans to implement the technology
in 2019/2020, and a further 21% are considering using it but have no current plans for
implementation.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
III
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 Executive Summary
Introduction
Companies of all sizes continue to invest heavily in data platforms technologies – and for good
reason. Data, if managed and leveraged correctly, can be the great enabler for many enterprises.
Efforts to better leverage data are often the impetus behind digital transforms initiatives, which
continue to be a primary driver for data platform technology adoption. But digital transformation
is just one aspect of the what is driving data platforms adoption. Enterprises have learned to
avoid falling too far behind on the technology curve and have shown a desire to continually
update their data platforms with pinpointed technologies. For instance, enterprises might add
caching to a database, blockchain or streaming technology to augment existing systems. In
essence, digital transformation initiatives may jumpstart a larger enterprise initiative, but it is the
ongoing updating, tweaking, and upgrading with data platform technologies that continue to fuel
the overall sector.

The data platforms sector is a subset of the broader Data, AI & Analytics channel at 451 Research.
As a whole, data platforms consist of all of the technologies that companies use to store, process
Data, if managed and analyze data relevant to their businesses. These technologies drive both internal- and
and leveraged external-facing business applications that can collect and analyze the data gathered from the
correctly, can business’s operations, which is then used to identify opportunities for improvement. However,
be the great data also comes from emerging sources such as social media and IoT sources. More specifically,
enabler for many data platforms consist predominantly of operational and analytic databases, including blockchain
enterprises. and distributed ledger technologies (DLT), but also distributed data-processing frameworks (aka,
Hadoop/Spark), data grid/cache and event/stream processing technologies.

The combined data platforms and analytics market generated revenue of $62.4bn in 2018,
excluding blockchain and distributed ledger technology (DLT), according to 451 Research’s
Data, AI & Analytics Market Monitor, September 2019 – and it is growing, thanks in part to the
current focus on what 451 Research calls ‘Pervasive Intelligence.’ This concept concerns the
combination of traditional and IoT data with big-data storage and processing, as well as artificial
intelligence, machine learning and other advanced analytics approaches. The market continues
to evolve, with new data platforms, products and services emerging to more efficiently store and
process data and provide access for those who need it.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
IV
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
Methodology
451 Research Market Maps™ are designed to provide a view of the vendor landscape by major
segment. The map highlights companies competing in multiple segments by connecting them
through a circuit line. Identification and placement of companies into these segments is based
on analysis, both published and unpublished, performed by 451 Research. This analysis includes
interviews, reports and advisory work with several thousand enterprises, vendors, service
providers and investors annually. 451 Research Market Maps™ are not intended to represent a
comprehensive list of every vendor operating in this market. Inclusion on 451 Research Market
Maps™ does not imply that a given vendor will be specifically featured in one or more 451
Research reports.

Reports such as this one represent a holistic perspective on key emerging markets in the
enterprise IT space. These markets evolve quickly, though, so 451 Research offers additional
services that provide critical marketplace updates. These updated reports and perspectives
are presented on a daily basis via the company’s core intelligence service, 451 Research Market
Insight. Forward-looking M&A analysis and perspectives on strategic acquisitions and the
liquidity environment for technology companies are also updated regularly via Market Insight,
which is backed by the industry-leading 451 Research M&A KnowledgeBase.

Emerging technologies and markets are covered in 451 Research channels including Applied
Infrastructure & DevOps; Cloud Transformation; Customer Experience & Commerce; Data, AI
& Analytics; Datacenter Services & Infrastructure; Information Security; Internet of Things;
Managed Services & Hosting; and Workforce Productivity & Collaboration.

Beyond that, 451 Research has a robust set of quantitative insights covered in products such
as Voice of the Enterprise (VotE), Voice of the Connected User Landscape, Voice of the Service
Provider, Cloud Price Index, Market Monitor, the M&A KnowledgeBase and the Datacenter
KnowledgeBase.

All of these 451 Research services, which are accessible via the web, provide critical and timely
analysis specifically focused on the business of enterprise IT innovation.

For more information about 451 Research, please go to: www.451research.com.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
V
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
Table of Contents
1. The 451 Take 1

2. The Data Platforms Landscape 2

Figure 1: 451 Research Data Platforms Market Map™����������������������������������������������������������������������������������� 3
Relational Operational Databases���������������������������������������������������������������������������������������������������������������������� 4
Non-Relational Operational Databases����������������������������������������������������������������������������������������������������������� 5
Analytic Databases����������������������������������������������������������������������������������������������������������������������������������������������������� 6
Distributed Data Processing Frameworks ������������������������������������������������������������������������������������������������������7
Event/Stream Processing���������������������������������������������������������������������������������������������������������������������������������������� 8
Distributed Data Grid/Cache���������������������������������������������������������������������������������������������������������������������������������� 9
Enterprise Blockchain��������������������������������������������������������������������������������������������������������������������������������������������� 10
Figure 2: Data Platforms Market Challenges and Innovations������������������������������������������������������������������� 11

3. Data Platforms Adjust to Continued Cloud Adoption 14

Figure 3. Enterprises Are Planning To Adopt Traditional Data Platforms in Cloud Environ-
ments��������������������������������������������������������������������������������������������������������������������������������������������������������������������������������������� 15
Figure 4: Major Market Players: Key Attributes������������������������������������������������������������������������������������������������ 16

4. Conclusions 22

5. Further Reading  23

6. Index of Companies 24

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
VI
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
1. The 451 Take
As expected, the data platforms market experienced a good deal of activity this past year. There
were several acquisitions. Two vendors were particularly active: HPE and Cloudera, with HPE
picking up three (Cray, BlueData, MapR) while Cloudera pulled in two (Hortonworks and Arcadia
Data). In terms of the distributed data-processing frameworks market involving MapR and
Cloudera, the acquisitions reflect a shifting market to the cloud, where it is expected that the
market will continue to show double-digit CAGR (24%) for the next five years. Along with the
acquisitions, several companies changed names. Data Artisans is now Ververica; MapD is now
OmniSci; and BlazingDB is now BlazingSQL. Also, a handful of new vendors have been added to
the market map as the evolving data platforms and analytics sector continues to spur innovative
new startups.

While these changes reflect an ever-shifting data platforms market, 451 Research forecasted
numbers continue to show a market with a strong growth trajectory. That is, the combined
data platforms and analytics markets revenue estimates, according to 451 Research Data, AI
& Analytics Market Monitor for September 2019, reveal a combined CAGR of 7.2% (excluding
blockchain and DLT). The fastest growing subsectors within the data platforms and analytics
segment between now and 2022 include distributed data processing, distributed data grid/cache
vendors, non-relational databases, event/stream processing, analytic platforms and relational
operational databases.

As highlighted in the 2017 and 2018 451 Research Data Platforms Market Map reports, cloud
adoption has been strong and will certainly continue. On-premises deployments, for instance,
which now reside at 49% for all data platforms, is expected to drop to 35% in two years. That
shortfall in on-premises deployments will be spread out over multiple iterations of cloud
environments (both public and private). But not all sectors will experience the same adoption
growth, suggesting a much more nuanced cloud adoption trend within the data platforms segment.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
1
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
2. The Data Platforms Landscape
The data platforms market can be split into five major categories (although there is considerable
overlap between them in some areas):

• Operational databases, including relational, non-relational, NoSQL and NewSQL databases.

• Analytical data platforms, which cover Apache Hadoop, Apache Spark, analytic databases and
data warehouses.

• Event/stream processing.

• Distributed data grid/cache systems.

• Distributed data processing frameworks.

• Enterprise blockchain and distributed-ledger technology (DLT).

Data platforms and distributed data-processing frameworks continue to be one of the hottest
sectors of IT, primarily because the volume of data continues to grow rapidly. Enterprises are not
only interested in storing and querying the data; they are increasingly interested in incorporating
real-time data ingestion and analytics. It comes as no surprise, therefore, that several
technologies on the market, including several new and emerging vendors, are aimed at helping
companies store, process and analyze their data.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
2
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
Figure 1: 451 Research Data Platforms Market Map™
Source: 451 Research, 2019

Non-Relational Operational
Databases
8K DATA SEQUOIADB
AEROSPIKE SPARSITY TECHNOLOGIES
ARANGODB STARCOUNTER
BITNINE TARANTOOL
DATASTAX TIGERGRAPH Distributed Data Processing
DGRAPH TIMESCALE
FAUNADB
Frameworks
FRANZ ALIBABA CLOUD ALIBABA CLOUD BLAZINGSQL
INTELLIGENT VIEWS AWS AWS CAZENA
KOVERSE IBM IBM CENTURYLINK
LUCIDWORKS
MARKLOGIC
MICROSOFT
ORACLE
MICROSOFT
ORACLE
JETHRO
PEPPERDATA
Enterprise Blockchain
MCOBJECT SAP SAP PHEMI ACCENTURE HYPERLEDGER
MEMGRAPH RACKSPACE RACKSPACE ROCKSET APPLIED BLOCKCHAIN INTEL
MONGODB CLOUDERA CLOUDERA SPLUNK BITFURY INTERLEDGER
NEO4J INSTACLUSTR INSTACLUSTR STARBURST DATA BLOCKAPPS JP MORGAN (QUORUM)
OBJECTIVITY INC ACTIAN ACTIAN STRATIO BLOCKCHAIN TECHNOLOGY PARTNERS MONAX
OBJECTLABS GOOGLE GOOGLE TRANSWARP BLOCKCYPHER MULTICHAIN
ONTOTEXT HPE HPE BLOCKSTACK R3
PRISMA BIGCHAINDB GIGASPACES CHAIN RIPPLE
QUASARDB FLUREE CISCO SWIRLDS
REDISGREEN VERVERICA TERADATA CITIZENS RESERVE THE IOTA FOUNDATION
SCALEGRID INTERSYSTEMS ESGYN QUBOLE CONSENSYS TRUSTED IOT ALLIANCE
SCYLLADB LEVYX SPLICE MACHINE DELOITTE VMWARE
DRAGONCHAIN
FAIRCOM SOFTWARE AG DATABRICKS ENTERPRISE ETHEREUM ALLIANCE ALIBABA CLOUD
HARPERDB FILAMENT AWS
PERCONA PIVOTAL FLEXVPC IBM
RAVENDB COUCHBASE TIBCO GOSPEL TECHNOLOGY MICROSOFT
YUGABYTE REDIS LABS ORACLE
BIGCHAINDB SAP
FLUREE
HPE

Event/Stream Analytic Databases

INTERSYSTEMS QUBOLE
Processing LEVYX TERADATA
Distributed Data Grid/Cache
60EAST VERVERICA
CLOUDKARAFKA DATABRICKS DATABRICKS 1010DATA GIGASPACES ALIBABA CLOUD
CONFLUENT ALTINITY AWS
FASTDATA.IO ALIBABA CLOUD ALIBABA CLOUD BLOBCITY COUCHBASE IBM
GUAVUS AWS AWS BRYTLYT REDIS LABS MICROSOFT
IMPETUS TECHNOLOGIES IBM IBM CITUS DATA ORACLE
INFLUXDATA MICROSOFT MICROSOFT CRATE.IO ALACHISOFT SAP
JKOOL ORACLE ORACLE EXASOL GRIDGAIN SYSTEMS
LIGHTBEND SAP SAP HETERODB HAZELCAST SOFTWARE AG
SAS INSTITUTE CLOUDERA CLOUDERA HPCC SYSTEMS MEMCACHIER PIVOTAL
STREAMLIO INSTACLUSTR INSTACLUSTR KINETICA RED HAT TIBCO
TIBCO TIBCO KOGNITIO SCALEOUT SOFTWARE
GOOGLE GOOGLE KX SYSTEMS
HPE HPE MICRO FOCUS
OMNISCI
SOFTWARE AG ACTIAN PARADIGM4
RYFT
SLICING DICE
Relational Operational SNOWFLAKE COMPUTING
SQREAM
Databases XTREMEDATA
YUGABYTE IGNITE TECHNOLOGIES IGNITE TECHNOLOGIES YELLOWBRICK DATA
RAVENDB MARIADB MARIADB
PERCONA MEMSQL MEMSQL
HARPERDB PINGCAP PINGCAP
FAIRCOM
PIVOTAL PIVOTAL
AIVEN
ALTIBASE RACKSPACE
CLUSTRIX ESGYN
COCKROACH LABS SPLICE MACHINE
CODERSHIP
COGNITECT ALIBABA CLOUD
CRUNCHY DATA AWS
EDGE INTELLIGENCE IBM
ENTERPRISEDB MICROSOFT
FAUNA ORACLE
GEMSERVERS SAP
HEROKU GOOGLE
LEANXCALE ACTIAN
NUODB
RDX
SWARM64
TMAXSOFT
VOLTDB

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
3
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 TIBCO
GOOGLE
HPE

SOFTWARE AG

Relational Operational Databases

Relational databases support and drive operational Relational Operational
business applications, including hybrid operational and Databases
analytic processing (HOAP) workloads.
YUGABYTE IGNITE TECHNOLOGIES
RAVENDB MARIADB
For the better part of the last five decades, relational
PERCONA MEMSQL
databases have dominated the overall database market.
HARPERDB PINGCAP
This category includes the databases used to support and
FAIRCOM
drive operational business applications (both internal-
PIVOTAL
and external-facing) by storing customer, product and
AIVEN
supplier records, as well as keeping track of day-to-day
ALTIBASE RACKSPACE
business transactions. Analytics is increasingly being
CLUSTRIX ESGYN
carried out on operational transactions, in what 451
COCKROACH LABS SPLICE MACHINE
Research refers to as HOAP. While hybrid workloads are
CODERSHIP
showing up in traditional relational systems, a growing
COGNITECT ALIBABA CLOUD
breed of relational-centered databases called NewSQL
CRUNCHY DATA AWS
(a term coined by 451 Research) has been particularly
EDGE INTELLIGENCE IBM
focused on HOAP workloads, often promoting scalability
ENTERPRISEDB MICROSOFT
while also maintaining atomicity, centricity, isolation and
FAUNA ORACLE
durability (ACID) in transactions.
GEMSERVERS SAP
HEROKU GOOGLE
LEANXCALE ACTIAN
NUODB
RDX
SWARM64
TMAXSOFT
VOLTDB

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
4
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
Non-Relational Operational Databases
Non-Relational Operational
Non-relational databases, which include NoSQL databases,
support and drive operational business applications. Databases
8K DATA SEQUOIADB
This category includes non-relational databases such as
AEROSPIKE SPARSITY TECHNOLOGIES
hierarchical, object and XML databases. NoSQL databases,
ARANGODB STARCOUNTER
which consist of multiple data models, including key-
BITNINE TARANTOOL
value, wide-column, document and graph data, represent
DATASTAX TIGERGRAPH
a significant component of this category. While NoSQL
DGRAPH TIMESCALE
is defined by four data models (key-value, wide-column,
FAUNADB
document and graph), many vendors support more than
FRANZ ALIBABA CLOUD
one data model, which are often referred to as multi-model
INTELLIGENT VIEWS AWS
NoSQL databases. Graph processing continues to see
KOVERSE IBM
growth. Several NoSQL databases now offer embedded
LUCIDWORKS MICROSOFT
or integrated graph capabilities within their database
MARKLOGIC ORACLE
systems, while there are now numerous stand-alone graph
MCOBJECT SAP
databases available in the market, both as on-premises
MEMGRAPH RACKSPACE
offerings and as cloud services.
MONGODB CLOUDERA
NEO4J INSTACLUSTR
OBJECTIVITY INC ACTIAN
OBJECTLABS GOOGLE
ONTOTEXT HPE
PRISMA BIGCHAINDB
QUASARDB FLUREE
REDISGREEN
SCALEGRID INTERSYSTEMS
SCYLLADB LEVYX

FAIRCOM SOFTWARE AG
HARPERDB
PERCONA
RAVENDB COUCHBASE
YUGABYTE REDIS LABS

Event/Stream
Processing
60EAST VERVERICA
CLOUDKARAFKA DATABRICKS
CONFLUENT
FASTDATA.IO ALIBABA CLOUD
D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
GUAVUS 5
AWS
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
IMPETUS TECHNOLOGIES IBM
INFLUXDATA MICROSOFT
JKOOL ORACLE
 REDIS LABS

Analytic Databases
nt/StreamAnalytic databases support and drive analytical Analytic Databases
INTERSYSTEMS QUBOLE
ocessing workloads, including data warehouses and data marts LEVYX TERADATA
that run analytical tools and applications, such as AI-
VERVERICA
driven analysis involving machine learning.
DATABRICKS DATABRICKS 1010DATA
ALTINITY
The analytic databases category includes analytical
ALIBABA CLOUD ALIBABA CLOUD BLOBCITY
databases, GPU-oriented database systems and other
AWS AWS BRYTLYT
platforms that make use of hardware acceleration. The
ES IBM IBM CITUS DATA
key theme among these systems and platforms is that
MICROSOFT MICROSOFT CRATE.IO
they are geared specifically toward carrying out analytics
ORACLE ORACLE EXASOL
and analytical processes, which are fundamentally
SAP SAP HETERODB
different from operational databases that allow
CLOUDERA CLOUDERA HPCC SYSTEMS
organizations to take the pulse of their businesses. Many
INSTACLUSTR INSTACLUSTR KINETICA
organizations would prefer not to burden an operational
TIBCO TIBCO KOGNITIO
database with the additional load for fear of slowing down
GOOGLE GOOGLE KX SYSTEMS
critical transactions. For that reason, analytics systems
HPE HPE MICRO FOCUS
and platforms are often, although not exclusively, used
OMNISCI
alongside their transactional counterparts. They are fed
SOFTWARE AG ACTIAN PARADIGM4
with copies of the data and can be queried more rapidly
RYFT
without the risk of affecting transactional performance.
SLICING DICE
That said, some hybrid systems have started to emerge
nal Operational
that allow transactional and analytical processing to
SNOWFLAKE COMPUTING
SQREAM
atabases occur in the same system.
XTREMEDATA
IGNITE TECHNOLOGIES IGNITE TECHNOLOGIES YELLOWBRICK DATA
MARIADB MARIADB
MEMSQL MEMSQL
PINGCAP PINGCAP

PIVOTAL PIVOTAL

RACKSPACE
ESGYN
SPLICE MACHINE

ALIBABA CLOUD
AWS
IBM
MICROSOFT
ORACLE
SAP
GOOGLE
ACTIAN

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
6
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
tabases
SEQUOIADB
SPARSITY TECHNOLOGIES
STARCOUNTER
TARANTOOL
Distributed
TIGERGRAPH Data Processing Frameworks Distributed Data Processing
Distributed
TIMESCALEdata processing frameworks support and
drive a variety of data management scenarios, from
Frameworks
analytics to operations, and often leverage open source
ALIBABA CLOUD ALIBABA CLOUD BLAZINGSQL
software AWS
such as Apache Hadoop, Apache Spark and AWS CAZENA
other emerging
IBM projects. IBM CENTURYLINK
MICROSOFT MICROSOFT JETHRO
The distributed
ORACLE data processing frameworks category ORACLE PEPPERDATA
combinesSAP the querying, compute and/or storage as SAP PHEMI
either separate open source or proprietary products.
RACKSPACE RACKSPACE ROCKSET
Providers in this category have a very different approach
CLOUDERA CLOUDERA SPLUNK
from that of traditional data warehousing or analytics
INSTACLUSTR INSTACLUSTR STARBURST DATA
database vendors, where all of the components are
ACTIAN ACTIAN STRATIO
natively integrated as part of a database management
GOOGLE GOOGLE TRANSWARP
system architecture.
HPE Vendors in this category leverage HPE
open source tools such as Apache Hadoop, Spark, or
BIGCHAINDB GIGASPACES
other open source processing or querying engines to
FLUREE
enable distributed data processing geared specifically for VERVERICA TERADATA
analytical workloads.
INTERSYSTEMS ESGYN QUBOLE
LEVYX SPLICE MACHINE

SOFTWARE AG DATABRICKS

PIVOTAL
COUCHBASE TIBCO
REDIS LABS

nt/Stream Analytic Databases

INTERSYSTEMS QUBOLE
ocessing LEVYX TERADATA
VERVERICA
DATABRICKS DATABRICKS 1010DATA
ALTINITY
ALIBABA CLOUD ALIBABA CLOUD BLOBCITY
AWS AWS BRYTLYT
ES IBM IBM CITUS DATA
MICROSOFT MICROSOFT CRATE.IO
ORACLE ORACLE EXASOL
SAP SAP HETERODB
CLOUDERA CLOUDERA HPCC SYSTEMS
INSTACLUSTR INSTACLUSTR KINETICA
TIBCO D a t a P l a t f o r m s M a r k e t M a pTIBCO
2019 KOGNITIO
7
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
GOOGLE GOOGLE KX SYSTEMS
HPE HPE MICRO FOCUS
OMNISCI
ACTIAN PARADIGM4
Event/Stream Processing
Event/Stream
Event/stream processing enables enterprises to make
decisions and automate actions based on the processing Processing
60EAST VERVERICA
and analysis of streams of event data from devices
CLOUDKARAFKA DATABRICKS
and applications. It includes publish-and-subscribe
CONFLUENT
messaging systems as well as full-fledged stream
FASTDATA.IO ALIBABA CLOUD
processing technologies.
GUAVUS AWS
Stream processing has been a part of the wider data IMPETUS TECHNOLOGIES IBM
and analytics market for several decades, but it was INFLUXDATA MICROSOFT
only in the last five-to-ten years that it emerged beyond JKOOL ORACLE
niche status. Event/stream processing enjoyed early LIGHTBEND SAP
success in markets with high-performance, real-time SAS INSTITUTE CLOUDERA
data processing requirements, such as financial services STREAMLIO INSTACLUSTR
and telecommunications. However, early products and TIBCO
services required specialist skills as well as use cases GOOGLE
that could justify a high price tag. Despite success in HPE
areas such as algorithmic trading, breakthrough vendors
struggled to make headway in more mainstream markets. SOFTWARE AG
Adoption of stream processing has escalated in the past
decade, however, driven by more widespread recognition
of the need for real-time data processing in industries Relational Operational
such as retail and advertising, as well as the emergence
of machine-to-machine data processing and IoT. Another
Databases
key factor has been the emergence of a slew of open YUGABYTE IGNITE TECHNOLOGIES

source messaging and stream processing projects, RAVENDB MARIADB

which have lowered the barriers to the development and PERCONA MEMSQL

deployment of stream processing applications. HARPERDB PINGCAP

FAIRCOM
PIVOTAL
AIVEN
ALTIBASE RACKSPACE
CLUSTRIX ESGYN
COCKROACH LABS SPLICE MACHINE
CODERSHIP
COGNITECT ALIBABA CLOUD
CRUNCHY DATA AWS
EDGE INTELLIGENCE IBM
ENTERPRISEDB MICROSOFT
FAUNA ORACLE
GEMSERVERS SAP
HEROKU GOOGLE
LEANXCALE ACTIAN
NUODB
RDX
SWARM64
D a t a P l a t f o r m s M a r k e t M a pTMAXSOFT
2019
8
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I VOLTDB
G H T S R E S E R V E D.
 FLUREE
HPE

Analytic Databases
INTERSYSTEMS QUBOLE
Distributed
LEVYX Data Grid/CacheTERADATA
In-memory distributed data layers – including data
Distributed Data Grid/Cache
DATABRICKS
caching 1010DATA
and distributed data grid processing – are used GIGASPACES ALIBABA CLOUD

to improve the performance of dynamicALTINITY

applications, AWS
ALIBABA
reduce CLOUD load and increasingly BLOBCITY
database act as primary COUCHBASE IBM
AWS
data BRYTLYT
serving layers for real-time applications. REDIS LABS MICROSOFT
IBM CITUS DATA ORACLE
Data caching, in its simplest form, creates
MICROSOFT an in-memory
CRATE.IO ALACHISOFT SAP
layer between the application and the database
ORACLE EXASOL that GRIDGAIN SYSTEMS
stores
SAP a cache of the most recently accessed
HETERODBdata, which HAZELCAST SOFTWARE AG
can then be queried to speed up the
CLOUDERA performance
HPCC SYSTEMS of MEMCACHIER PIVOTAL
dynamic applications by alleviating the KINETICA
INSTACLUSTR database load. RED HAT TIBCO
In the
TIBCO2000s, a group of startups emerged with data
KOGNITIO SCALEOUT SOFTWARE
cache products to serve enterprise low-latency
GOOGLE KX SYSTEMS data
access
HPE and query needs. Niche low-latency requirements,
MICRO FOCUS
such as algorithmic trading in financial services,
OMNISCI are no
longer
ACTIANthe only ones to require an in-memory
PARADIGM4approach
– more and more applications are latency-sensitive.
RYFT Over
time, vendors in this space have added distributed
SLICING DICE grid
processing capabilities, making theirCOMPUTING
SNOWFLAKE products suitable
for storing and processing – as well as querying
SQREAM – data
in memory, as well as improved reliability capabilities
XTREMEDATA
to IGNITE
support persistence. While these
TECHNOLOGIES products
YELLOWBRICK DATAare not
databases
MARIADB per se, we have seen some large enterprises
standardize
MEMSQL on distributed data-grid products as their
primary
PINGCAPdata processing tiers, relegating the database to
the role of persistent storage.
PIVOTAL

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
9
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 Distributed Data Processing
Frameworks
ALIBABA CLOUD BLAZINGSQL
AWS CAZENA
IBM CENTURYLINK
Enterprise Blockchain
MICROSOFT
PEPPERDATA
ORACLE and DLT are cryptographically
JETHRO
Enterprise Blockchain
Blockchain enabled
SAP
distributed PHEMI over data
database systems in which the control ACCENTURE HYPERLEDGER
RACKSPACE ROCKSET
evolution (read/write) is shared among participating entities, APPLIED BLOCKCHAIN INTEL

transactions that are executed basedSPLUNK

andCLOUDERA on consensus. BITFURY INTERLEDGER
INSTACLUSTR STARBURST DATA BLOCKAPPS JP MORGAN (QUORUM)
Enterprise
ACTIAN blockchain and DLT is a subsegmentSTRATIO of BLOCKCHAIN TECHNOLOGY PARTNERS MONAX
the overall blockchain and DLT market. It is made up
GOOGLE TRANSWARP BLOCKCYPHER MULTICHAIN
of organizations
HPE developing and driving open source BLOCKSTACK R3
or proprietary blockchain and DLT architecture,
GIGASPACES CHAIN RIPPLE
protocols and platforms, as well as vendor offerings CISCO SWIRLDS
(e.g., blockchain as a service) that caterTERADATA
VERVERICA to enterprise CITIZENS RESERVE THE IOTA FOUNDATION
requirements,
ESGYN be it general-purpose or vertical-specific.
QUBOLE CONSENSYS TRUSTED IOT ALLIANCE
The enterprise blockchain and DLT landscape is still
SPLICE MACHINE DELOITTE VMWARE
fragmented and multi-protocol. While developers DRAGONCHAIN
continue to innovate, industry consortia drive not only
DATABRICKS ENTERPRISE ETHEREUM ALLIANCE ALIBABA CLOUD
the evolution of blockchain protocols, but also their FILAMENT AWS
adoption
PIVOTAL across a wide set of industries. Blockchain is a FLEXVPC IBM
multiparty
TIBCO collaborative effort, not only in its use but also GOSPEL TECHNOLOGY MICROSOFT
in its development. While blockchains are append-only ORACLE
with transactions chronologically chained together in BIGCHAINDB SAP
blocks, other DLT systems may use other data structures. FLUREE
Blockchain technology has become part of the enterprise HPE
digital portfolio, with 50% of organizations at least
considering its implementation and 10% reporting it as a
top IT priority for 2019.
Analytic Databases
INTERSYSTEMS
The QUBOLE
blockchain-as-a-service (BaaS) market has
LEVYX
significantly expanded in the last year,TERADATA
seeking to ease
adoption of blockchain technology for enterprises, with
Distributed Data Grid/Cache
DATABRICKS
a mission 1010DATA
to hide infrastructure complexity, fill in skills GIGASPACES ALIBABA CLOUD
ALTINITY
gaps, improve cost and provide customers with choice. AWS
TheALIBABA
BaaSCLOUD
space is getting more crowded BLOBCITY
with not only COUCHBASE IBM
AWS BRYTLYT
cloud heavyweights, but also smaller providers. REDIS LABS MICROSOFT
IBM CITUS DATA ORACLE
MICROSOFT CRATE.IO ALACHISOFT SAP
ORACLE EXASOL GRIDGAIN SYSTEMS
SAP HETERODB HAZELCAST SOFTWARE AG
CLOUDERA HPCC SYSTEMS MEMCACHIER PIVOTAL
INSTACLUSTR KINETICA RED HAT TIBCO
TIBCO KOGNITIO SCALEOUT SOFTWARE
GOOGLE KX SYSTEMS
HPE MICRO FOCUS
OMNISCI
ACTIAN PARADIGM4
RYFT
D a t a SLICING
P l a t fDICE
o r m s M a r ket M a p 2 0 1 9
10
© C O P Y RCOMPUTING
SNOWFLAKE I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.

SQREAM
XTREMEDATA
IGNITE TECHNOLOGIES YELLOWBRICK DATA
Figure 2: Data Platforms Market Challenges and Innovations
Source: 451 Research, 2019

MARKET SEGMENT KEY CHALLENGES INNOVATIONS

RELATIONAL Traditional relational databases require that Relational databases are addressing flexible data
OPERATIONAL the data reside in a certain structure (database models. Nearly all of the major relational database
DATABASES schema), which is established up front before management system (RDBMS) vendors enable
loading the data. Relational databases can be the ability to store data in JSON, a data format
highly optimized when the schema is aligned often associated with NoSQL databases. JSON
with usage scenarios. However, usage scenarios data provides flexible schema on read approaches
can and do change, which requires significant versus a schema on write approach.
effort to develop new schemas, possibly
Transactional and analytical processing has
resulting in performance challenges.
historically been separated, but several vendors
Operational databases function well now provide functionally or specifically designed
for transactional workloads, but many systems that address hybrid workloads. NewSQL
organizations also carry out analytics on vendors are targeting this area, as are many
that data, oftentimes referred to as mixed or traditional relational systems that address hybrid
hybrid workloads. The traditional approach to processing by serving the analytics in memory
accommodating both transactions and analytics while keeping transactions on disk, for instance.
has been to incorporate separate systems As such, hybrid systems introduce a variety of
to prevent analytical queries from affecting new use cases when analytics can be integrated
incoming transactions and vice versa. with incoming transactions.

NON-RELATIONAL NoSQL is defined as four different data Providing multiple data models (called multi-
OPERATIONAL models (key-value, document, wide-column model) continues to be common for many
DATABASES and graph) such that there are virtually no NoSQL databases. Vendors continue to provide
common standards across the NoSQL family efficiencies between data models, such as
of databases. NoSQL vendors tend to leverage providing data querying among the different
and employ different query languages, security models without needing to copy the data from
features, and scaling approaches and data one model to another.
consistency standards.
NoSQL databases continue to play a prominent
As distributed systems with non-relational data role in helping organizations address digital
models, NoSQL databases often emphasize transformation efforts, where globally distributed
scalability benefits that can accommodate high applications are common. As such, vendors
data and user concurrency loads. However, are targeting hybrid workloads (operational
depending on the type of workload, data transactions and analytics), mobile applications,
consistency comes into play. ACID compliance and IoT scenarios.
has long been table stakes for relational
databases but can vary between NoSQL
databases, with some providing full ACID
while others offer what has come to be called
‘eventual consistency.’

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
11
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
ANALYTIC DATABASES NoSQL is defined as four different data Providing multiple data models (called multi-
models (key-value, document, wide-column model) continues to be common for many
and graph) such that there are virtually no NoSQL databases. Vendors continue to provide
common standards across the NoSQL family efficiencies between data models, such as
of databases. NoSQL vendors tend to leverage providing data querying among the different
and employ different query languages, security models without needing to copy the data from
features, and scaling approaches and data one model to another.
consistency standards.
NoSQL databases continue to play a prominent
As distributed systems with non-relational data role in helping organizations address digital
models, NoSQL databases often emphasize transformation efforts, where globally distributed
scalability benefits that can accommodate high applications are common. As such, vendors
data and user concurrency loads. However, are targeting hybrid workloads (operational
depending on the type of workload, data transactions and analytics), mobile applications,
consistency comes into play. ACID compliance and IoT scenarios.
has long been table stakes for relational
databases but can vary between NoSQL
databases, with some providing full ACID
while others offer what has come to be called
‘eventual consistency.’

DISTRIBUTED Data warehouse systems have traditionally been According to VotE: Data and Anlytics data, while
DATA PROCESSING on-premises deployments that often receive the current focus is on on-premises deployment,
FRAMEWORKS data from operational database systems. With the highest demand from enterprises in terms of
data volumes increasing, organizations can future adoption is for as-a-service deployments,
have challenges accommodating not only the especially standard database platforms and
data but also the need to scale the system to tools: 27% of respondents say they plan to adopt
accommodate more user concurrency. analytic database software as a service over the
next two years and 25% plan to do the same with
With organizations looking to be more analytics-
relational operational database software.
driven, that is putting increased pressure on the
type of analytics that can be carried out. Many In-database machine learning, which allows
enterprises are wanting to adopt of machine organizations to develop machine learning models
learning to help drive business decisions directly within the database and then to deploy
yet carrying out machine learning requires them to specific applications, is seeing increased
organizations to implement separate systems for adoption. The idea of running machine learning
that technology, often operated by highly skilled in a data warehouse should not necessarily be
data scientists. a foreign concept because data warehouse
systems are meant to support businesses through
data analysis.

EVENT/STREAM Real-time data processing has traditionally Traditional complex event processing
PROCESSING been expensive, and therefore reserved for technologies have required a specific set of
high-value applications. The emergence of skills to program and query. Recent innovations
numerous open source stream processing in stream processing include support for SQL
engines has made the technology mainstream, queries – reducing the barriers to adoption and
but enterprises still need to understand the the delivery of business value.
potential opportunities and develop applications
A greater focus on managing and automating
to take advantage of them.
the flow of data throughout the organization
The stream processing market is particularly (rather from simply point A to point B) elevates
challenged in relation to the potential for the value of stream processing to make it more
confusion with several projects that are hard of a strategic imperative to support the concept
for non-experts to differentiate, and others that of the event-driven enterprise.
overlap considerably but can be complementary.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
12
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
DISTRIBUTED DATA Although adoption of in-memory data- In-memory data grid/cache has a key role to
GRID/CACHE processing technologies has become more play in delivering the level of responsiveness
widespread, vendors in this space still need to required to support hybrid operational and
educate the market on their approach and how analytics processing – enabling more rapid
it differentiates from the various offerings that analysis of data and faster business insight.
are out there. Enterprises are investing in data-processing
technology and analytics to become more data-
The increased integration of in-memory
and insight-driven, and HOAP workloads are of
processing capabilities into traditional
growing interest and importance.
databases has the potential to overshadow the
benefits of in-memory data grid/cache. There is potential for users to evolve data grid/
cache adoption into in-memory data layers that
relegate the role of the database to back-end
data storage. However, data grid/cache vendors
also concede that they come up against pure in-
memory database vendors in some competitive
situations. Data grid/cache providers are
expanding their offerings and becoming in-
memory computing vendors.

ENTERPRISE As is often the case amid the emergence of The enterprise blockchain and DLT market is
BLOCKCHAIN a new technology, there has been a lot of, maturing, driven by a growing focus on the right
sometimes irrational, exuberance around use cases, easier access to the technology,
blockchain (especially around cryptocurrencies), and a growing interest by businesses in
however, we believe that a growing focus on implementing it across industries. Also, as more
enterprise use cases will provide balance. data and insights in terms of value gained from
real-world implementations are revealed and
The blockchain landscape is still evolving – it’s
shared, adoption will accelerate.
technical and fragmented. There are several
protocols coexisting, with no clear winners. The blockchain-as-a-service BaaS space
The number of protocols has been on the rise is getting increasingly crowded, with both
because developers are continuing to innovate, cloud incumbents and smaller entrants with a
as there is no perfect solution that is able to mission to ease and accelerate the adoption
address all needs at once. of blockchain technology by enterprises.
Interoperability efforts will be key to have
blockchain networks connect with one another
and allow users to transact across them.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
13
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
3. Data Platforms Adjust to
Continued Cloud Adoption
The data platforms market is quite large with more than 300 companies that we track and
analyze as a subsector of the Data, AI & Analytics channel. While we always expect to see a
number of market changes from year to year among the vendors, this year was no different as
there were several vendor acquisitions and other events, highlighting some market trends.

One particular area – distributed data-processing frameworks such as Hadoop – experienced a

number of market changes, providing some insights to where this market is headed. The Hadoop
market first started to shift when Cloudera announced its intention to acquire fellow rival
Hortonworks in October 2018, with the deal officially closing in January 2019. Both companies
cited cost synergies and cross-selling opportunities as the primary driver for the deal, primarily
because Cloudera and Hortonworks had been competing directly with each other for some time
before the announcement. Fast forward to early summer, and Cloudera and MapR made news
when both companies saw unexpected revenue dips, albeit for different reasons. For Cloudera,
its revenue challenges were related to the timing of its acquisition of Hortonworks, which led to
customers delaying upgrades in response to forthcoming product announcements. For MapR,
the company reported that it had changed its go-to-market strategy from a direct sales model
to an indirect sales model, which materially impacted revenue. As a result of this, HPE acquired
MapR’s business assets in August 2019.

Certainly, the news from MapR and Cloudera did not bode well for the Hadoop market in general.
These events should not indicate a falling market because the market is much larger than two
vendors, and a falling market would indicate declining revenue in other areas as well. Hadoop-
based services (distributed data-processing frameworks) are also peddled by the large cloud
providers, notably Amazon Web Services (AWS) with its Amazon EMR service, Microsoft Azure
with its HDInsight service, and Google Cloud Platform with its Google Cloud Dataproc service.
There are also cloud-based, Hadoop-related services offered by Qubole, SAP’s Cloud Platform
Big Data Service, Oracle’s Big Data Cloud Service, IBM’s Analytics Engine service, HPE’s Blue
Data and Cazena.

Moreover, the events with MapR and Cloudera have signaled a shifting of the market, instead of a
falling market. Current data from 451 Research’s Market Monitor for Data Platforms & Analytics
suggests that the Hadoop market (what we call distributed data-processing frameworks) will
continue to show double-digit CAGR growth (24%) for the next five years. Both Cloudera and

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
14
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 MapR have historical roots in on-premises deployments, and while both were moving to increase
their cloud services, only Cloudera offered a cloud-managed service, with its Altus Cloud, while
MapR provided its products by way of IaaS deployments. As such, the challenges of Cloudera
and MapR were not so much an indication of a falling market, but a shifting market that is
transitioning to the cloud.

Overall, cloud adoption continues to be the most dominant trend within the data platforms
sector, and as described, has begun to impact other sectors such as the distributed data-
processing frameworks (Hadoop) market. While there are still organizations that have yet to
adopt a cloud strategy for their data platforms, there are many organizations that have already
transitioned to the cloud and are now expanding their efforts, suggesting that enterprise
angst over moving database systems to the cloud is waning. In fact, recent findings from 451
Research’s VotE: Data and Analytics reveal enterprises continue to adopt traditional data
platforms but are looking to do so as cloud services, as shown in Figure 3.

Figure 3. Enterprises Are Planning To Adopt Traditional Data Platforms in Cloud Environments
Source: 451 Research’s Voice of the Enterprise: Data and Analytics, 1H 2019
Q. Looking ahead, does your organization plan to begin using any of the following types of data platforms or services within
the next two years?

PLANNED ADOPTION OF DATA PLATFORMS

% of respondents (n=467)

Analytic database
14% 27%
(data warehouse) software

41%
Relational operational database 8% 25%

31%
Big data platform software
(e.g., Hadoop, Spark)
12% 17%

29%
NoSQL operational
13% 15%
database software
On-premises
28% As-a-service
Distributed ledger technology 12% 13%

25%

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
15
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 Collectively, enterprises continue to find value in traditional data platform technologies such as
analytic and relational databases, with planned adoption within the next two years at 41% and
31%, respectively. However, newer technologies such as big data software (Hadoop and Spark),
NoSQL, and distributed ledger systems will likewise see adoption, albeit slightly behind analytic
and relational databases. In all cases, however, adoption will be driven by cloud services, rather
than on-premises deployments, a trend we expend to continue upwards.

Figure 4: Major Market Players: Key Attributes

Source: 451 Research, 2019
The following table contains a list of the vendors that are featured in more than two of the Data Platforms Market Map
segments. This list is not designed to be exhaustive and only contains vendors competing in multiple segments.

COMPANY NAME KEY ATTRIBUTES

ACTIAN Database and data management vendor Actian positions itself around a hybrid data strategy,
which divides the company’s products into three primary areas: data integration, data management
and data analytics. The company’s data platforms products include Actian X (hybrid database),
NoSQL Object Database, and Zen Embedded Database (IoT database), Vector (analytic database),
Avalanche (cloud data warehouse).

ALIBABA CLOUD Alibaba Cloud is the dominant public cloud provider in China and is ramping up its effort in other
markets, including the US and Europe. The company already has a broad array of data platform
cloud service offerings, including multiple databases – both relational and non-relational, as well
as transactional and analytic – and also offers services addressing distributed data processing
frameworks, event and stream processing, distributed data grid/cache, and blockchain.

AWS AWS’ Redshift proved that there is demand for data warehouse as a service, but AWS has a large
range of data processing and storage offerings in the cloud, as well as addressing operational
databases, Hadoop, data caching and stream processing. AWS has been slowly and carefully dipping
its toes in the blockchain pond since 2016. At re:Invent 2018, AWS announced the launch of two
fully fledged offerings – Amazon Quantum Ledger Database (QLDB), a database with blockchain
characteristics (but centralized), and Amazon Managed Blockchain, a managed blockchain service
that allows users to create and manage blockchain networks using Ethereum and Hyperledger
Fabric.

BIGCHAINDB BigchainDB is essentially a database with blockchain characteristics. It combines the features of a
more conventional distributed database with blockchain technology, so multiple organizations can
share a common set of data in a decentralized way. BigchainDB has abstracted its core code from
the underlying database to be able to plug into multiple database back ends. It currently supports
RethinkDB and MongoDB database programs.

CLOUDERA Founded in 2008, Cloudera filed for an IPO in March 2017. In October 2018, it announced its
acquisition of Hortonworks, which officially closed in January. Before the acquisition, Cloudera and
Hortonworks were rivals of sorts, although both shared a collective interest in and commitment to
Apache Hadoop and related projects that would become known as the broader Hadoop ecosystem.
With two different product lines and go-to-market strategies, there were some things to sort out.
With product reconciliations being addressed with the company’s forthcoming Cloudera Data
Platform (CDP), which serves as a type of umbrella term for the company’s product offerings, there
was still needed resolution as to which open source model to adopt. That has been settled, with
Cloudera announcing that it will adopt a full, 100% open source model patterned after Red Hat’s.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
16
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
COUCHBASE Couchbase can be categorized as a multi-model NoSQL database. The company’ Couchbase Server,
which can also be used as a distributed cache, and Couchbase Mobile products enable the company
to position itself as a modern, general-purpose database that supports a variety of workloads from
mobile, Internet of Things and web-based applications. There is also the company’s SQL-based
query tool called N1QL (pronounced nickel), which is a declarative language that extends SQL for
JSON.

DATABRICKS Founded in 2013, the company is not only known for the being founded by the original creators of
the open source Apache Spark project, but it is also the commercial entity supporting it, and greatly
influences the project’s direction. The company offers its Unified Analytics Platforms for analytics
leveraging Spark and is available on AWS and Azure. Other services, which can integrate with the
Unified Analytics Platform, include Delta Lake that provides ACID transactions, MLflow that enables
ML model lifecycles management, Databricks Workspace for analytics collaboration, and Databricks
Runtime for Machine Learning that provides a scalable cluster environment for training ML and
deep learning models.

ESGYN Esgyn, an HP spinoff and commercial supporter of the open source Apache Trafodion project, offers
an RDBMS product that runs on Hadoop. The EsgynDB product, therefore, provides both OTLP
and OLAP processing within the same framework. The technology is based on HP’s NonStop SQL
database geared toward transactional processing. Recently, the company rolled out a managed
cloud service called Strato.

FAIRCOM Founded in 1979, FairCom provides a multi-model database that consists of both a SQL relational
database as well as a key-value NoSQL store. The company claims strong consistency noting
its ACID compliance, for which the company drives an OEM strategy targeting mission-critical
application use cases. The company also offers an IoT product based on the company’s legacy
technology.

FAUNADB Fauna was founded in 2012 but emerged in late 2018 with a generally available database called
FaunaDB. Referred to as a ‘relational SQL’ database, FaunaDB borrows characteristics from both
those database domains. Fauna is targeting enterprises that need a database to drive globally
distributed applications. FaunaDB has its own protocol that enables transaction consistency over
geographically dispersed datacenters, as well as its own query language and API platform.

FLUREE North Carolina-based Fluree has refocused its value proposition from blockchain database to
blockchain-backed data management platform for decentralized applications. Fluree’s technology
stack is now composed of two key components: FlureeDL, an immutable distributed ledger where
every update is secured with blockchain cryptography in chronological order, and FlureeDB, a graph
database for building applications on top of FlureeDL.

GIGASPACES GigaSpaces has a long track record of providing in-memory data-processing for data grid and cache
use cases. With the latest enhancements to its InsightEdge Platform, which combines the company’s
eXtreme Application Platform in-memory data grid/cache technology with its InsightEdge grid-
enabled distribution of Apache Spark, is aiming to further simplify and accelerate big-data processing
for ‘instant’ business insight.

GOOGLE While Google’s distributed data processing research inspired Hadoop as well as multiple NoSQL
projects, the company is still finding its feet in terms of its data-related cloud services. The portfolio
includes relational databases (Google Cloud SQL and Google Cloud Spanner), NoSQL (Google Bigtable
and Cloud Datastore), Hadoop (Google Dataproc), analytic database (Google BigQuery) and stream
processing (Cloud Dataflow). Google has partnered with BlockApps and Digital Asset to allow its cloud
customers to explore ways to use blockchain and DLT.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
17
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
HARPERDB Founded in early 2017, the goal behind HarperDB is to incorporate the scalability, ease of use, and
data flexibility of NoSQL, yet maintain its SQL compatibility for analytic workloads. In that sense,
the company offers neither a NoSQL (as defined by the four NoSQL data models) database nor an
SQL RDBMS. Instead, HarperDB refers to its offering as a SQL-NoSQL database. Given this, 451
Research’s database categorization would put it among the NewSQL family of databases, many of
which also support some NoSQL capabilities.

HPE HPE completed three acquisitions last year: BlueData, Cray and MapR. BlueData offered a
virtualized container-based platform for enterprises looking to power big-data analytical
workloads, including machine learning and deep learning. Best known for its supercomputers,
Cray’s product line included a data analytics platform product called Urika-GX that, in addition
to including supercomputing functionally, also comes with Apache Hadoop, Apache Spark and
the company’s graph-based engine. Once known solely for its commercial Hadoop distribution,
MapR offers the MapR Data Platform, which features analytical-based workloads, including AI and
ML, enabled by its propriety data layer consisting of a file system, NoSQL database and event-
processing engine. Further, HPE announced the launch of its Mission-Critical DLT offering at the
end of 2017.

IBM IBM remains a powerful force in the data platforms space, with products in almost all segments,
particularly as the company bets big on ‘cognitive computing’ with its Watson brand. The
company’s DB2 database is a mainstay in the industry and includes in-memory capabilities as
well as BLUAcceleration. The company has likewise been driving a strong cloud strategy with a
number of new services available on its Bluemix cloud platform. IBM is among the early believers
in blockchain technology and is a premier member and substantial contributor to the Hyperledger
project.

IGNITE TECHNOLOGIES Owned by ESW Capital, Ignite Technologies has made a number of acquisitions over recent
years to assemble a portfolio of database-related products and services. That portfolio includes
the Infobright analytic database, as well as the ObjectStore in-memory object database and the
ScaleArc database load balancing software.

INSTACLUSTR Instaclustr started out with a managed cloud service for Apache Cassandra but has since added
several other open source products, including Apache Spark, Lucene, Zeppelin, Elassandra for
Elastisearch and ScyllaDB, a NoSQL database pitched as a potential replacement for Cassandra.
In addition, Kafka for streaming was rolled out in May 2018. However, the company will now focus
primarily on four core service offerings: Apache Cassandra, Apache Spark, Apache Kafka and
Elasticsearch, which the company currently offers as part of an early access program, with a fully
managed service expected later.

INTERSYSTEMS InterSystems has been around for the better part of 41 years. Having been a mainstay in the
healthcare industry for some time, InterSystems has built a loyal following, but it has also gained
considerable traction in other sectors such as financial services, logistics, logistics/supply chain
and manufacturing. In 2017, InterSystems announced the IRIS Data Platform that was designed to
combine two of InterSystems’ products into one platform: Caché, a NoSQL-like object database,
and Ensemble, a rapid integration and development platform. The company pitches the IRIS Data
Platform as a highly flexible offering that can handle both transactional and analytical workloads
for data-heavy applications.

LEVYX Founded in 2013, Levyx is focused on high-performance distributed data processing, designed
specifically, to take advantage of Flash/NVMs. The portfolio includes the Helium key value store,
the Xenon data analytics processing engine, and the Radon, which combines Helium and Xenon
for exabyte-scale use cases.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
18
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
MARIADB MariaDB offers the MySQL-compatible X3 platform that combines MariaDB TX (transactions)
and Maria AX (analytics) that blends transactional and analytical processing while leveraging the
company’s pluggable storage engine approach. Cloud-wise, MariaDB had previously launched
a managed cloud service, but MariaDB has noted that a cloud-native and Kubernetes-friendly
offering called SkySQL will be arriving later in 2019.

MEMSQL Founded in 2011 and headquartered in San Francisco, MemSQL drives a type of distributed
SQL offering with its ability to scale its SQL-based database that is capable of handling ACID
transactions. The company promotes it ability to carry out high performant data ingest of varied
data types, which include relational as well as key-value, JSON, and timeseries data. Further,
MemSQL notes its ability to drive hybrid workloads – that is, workloads capable of handling
transactions as well as analytics within the same system.

MICROSOFT From Excel all the way up to large enterprise data warehouse technology, few other companies
can claim to have quite the breadth of technologies that Microsoft can and the company is a
key player in both the database and distributed data processing framework sectors. Many of its
technologies are also made available in its Azure cloud, which it is using to address emerging
segments such as machine learning, Hadoop and document database capabilities. Microsoft’s
fully managed blockchain service leveraging the permissioned Quorum protocol – Azure
Blockchain Service – is now available in a public preview mode.

ORACLE While a giant in the relational database market, Oracle has been making a deliberate and
aggressive push into the cloud space, pointing to Amazon AWS and Microsoft Azure as its new
primary competitors. With the focus on cloud, the company has expanded its cloud capabilities
through a mixture of acquisitions and organic development. Even though cloud is a strong focus,
not least for its Autonomous Database Service, Oracle continues to drive its database business
for the company’s ‘engineered systems’ of products that utilize hardware from its acquisition of
Sun Microsystems. Oracle’s blockchain platform is called Oracle Autonomous Blockchain Cloud
Service (OABCS) and was made generally available in 2018.

PERCONA Percona has always been about adopting leading open source databases and wrapping services
around them. The company follows a pure open source model in that any and all enterprise
functionality, including tooling, that it develops and integrates with these databases, it also makes
available as open source. Percona considers itself technology-agnostic with its main focus on
MySQL and MongoDB, although it does support other databases such as MariaDB.

PINGCAP PingCAP follows a modular design for its distributed SQL-based database known as TiDB,
which is provided as open source; the startup also provides some enterprise functionality. The
company’s TiDB database, which is built with a collection of components and offered as open
source, can handle globally consistent transactions (ACID), but the company points out that many
of its customers want the ability to do analytics on those transactions. Thus, TiDB leverages the
company’s TiSpark component (based on Apache Spark) to carry out analytics on the incoming
transactional data.

PIVOTAL Now being subsumed by VMware, Pivotal Software was created to build a commercial open
source business around assets spun out of EMC and VMware, including Cloud Foundry and, in the
big data market, the Pivotal Greenplum analytic database, Pivotal GemFire in-memory data grid/
cache, and Pivotal HDB for SQL-on-Hadoop.

QUBOLE Qubole was founded in 2011 and offers the Qubole Data Platform (also called Qubole Data
Service or QDS), which is a cloud-based data processing platform service that runs on Amazon
AWS, Microsoft Azure, Google Cloud Platform and Oracle Cloud. The service enables data
democratization while also enabling easy data management capabilities, both from a user
perspective (self-service) and an administrative perspective. The company takes a collective-
engines approach that leverages open source processing tools such as Apache Spark, Presto and
Hive.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
19
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
RACKSPACE Managed cloud computing specialist Rackspace offers a range of data platform hosted
services including Hadoop (Cloud Big Data Platform and Managed Big Data Platform); managed
MySQL databases; and managed MongoDB, Redis and Elasticsearch, leveraging the company’s
ObjectRocket cloud platform acquisition.

RAVENDB Founded in 2008 and based in Israel, RavenDB offers a multi-model, NoSQL database capable of
handling ACID transactions. As a NoSQL multi-model database purveyor, RavenDB is based on the
document model, but can also handle key-value as well as binary data. The company notes that its
database often gets deployed in two common ways. One is as a stand-alone cluster that can drive
a single application, taking advantage of the database’s multi-model capabilities. The other is as a
distributed cluster, such as in a point-of-sale scenario, which the company notes is a common use
case for many of its customers.

REDIS LABS Founded in 2011, Redis Labs is the largest and most recognized commercial supporter of the
open source Redis NoSQL database, an in-memory key-value database that can also function
as a cache, memory broker or primary database, as it can also persist data. The company’s core
product is a database-as-a-service offering, but the company also provides an on-premises,
enterprise-grade version of its database. Central to Redis Labs is its modules strategy that
was introduced in 2016 that effectively enables the company to provide native multi-model
operations.

SAP SAP remains best known for its enterprise application software, but it is also a significant player
in the data platforms space due to its investment in HANA as well as its acquisition of Sybase –
which brought capabilities in operational and analytic databases, as well as stream processing
that it continues to invest in via R&D. The company has recently focused on bringing the benefits
of HANA’s query execution and data modeling approaches to data stored in other environments
with the launch of SAP HANA Cloud Services.

SOFTWARE AG Software AG boosted its data platforms credibility with the 2013 acquisition of the Apama CEP
technology from Progress Software, but it already had a wide range of capabilities such as in-
memory data grid/cache (Terracotta) and relational database (Adabas/Natural) all of which are
now combined as part of its Data & Analytics Platform.

SPLICE MACHINE Splice Machine drives a Hadoop- and Spark-based RDBMS offering, capable of handling both
OLTP and OLAP workloads. Traditionally, organizations have shied away from dual workload
systems, but Splice Machine claims its optimization approach to segment queries doesn’t
negatively impact the differing workloads. Having started out offering proprietary software,
the company recently open sourced its database and has since ventured into the cloud with a
database-as-a-service offering. From positioning standpoint, Splice Machine not only combines
transactional and analytical workload but intelligence-based workloads as well by leveraging
machine learning.

TERADATA A core player in the data warehousing space, the firm makes around half of its revenue from
selling its warehouse appliances and the rest from associated software, services and consulting.
The company’s Teradata Vantage was a major rebranding of the Teradata Analytics Platform.
As the company’s flagship product, Vantage consists of three components: embedded analytics
engines and functions, integrated tools and languages, and the ability to accept varied data types.

TIBCO While still perhaps best known for its data integration capabilities, TIBCO has assembled a
growing portfolio of data platforms and analytics products and cloud services, including the
StreamBase stream processing engine, the ActiveSpaces in-memory data grid and its own
distribution of Apache Kafka, as well as the TIBCO Graph Database and TIBCO ComputeDB
combination of Apache Spark and Apache Geode.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
20
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
VERVERICA Previously known as Data Artisans, Ververica was renamed in early 2019 shortly after its
acquisition by Alibaba. The company remains focused on building a commercial business around
the Apache Flink project, which is designed to support both stream processing and batch
processing of historical data. Additional functionality offered in the Ververica Platform includes
application lifecycle management as well as streaming ledger, which offers support for ACID
transactions on streaming data.

YUGABYTE Based in Sunnyvale, California, YugaByte was founded in 2016 and provides a database (also
available as YugaByte Cloud) that blends the best of NoSQL with the best of SQL, specifically
targeting those organizations that want to deploy globally distributed applications. The company
targets a certain custom type, particularly those organizations that have a need to carry out
ACID transactions and to be highly performant while needing the ability to scale across globally
distributed geographies.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
21
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
4. Conclusions
With the distributed data-processing frameworks market seeing changes, we would expect
to see other sectors experience changes in the foreseeable future. Two markets might be
particularly susceptible. The analytics database market, which shows strong cloud adoption,
is likely to see changes based on this trend. The NoSQL database market is also likely to see
changes given the sheer number of vendors competing in the space. We wouldn’t be surprised
if a few NoSQL vendors filed for an IPO, given that this market is now 10 years old and showing
signs of maturity. Technology overlap between markets will likewise continue, especially
between the relational-operational databases and analytic databases segments, as enterprises
show an increased appetite for hybrid workloads.

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
22
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
5. Further Reading
451 Perspective: The tyranny of choice – are there too many stream processing engines?,
July 2019

451 Perspective: Service provider blockchain opportunities and strategies, July 2019

Blockchain adoption highlights, June 2019

State of the Enterprise Blockchain Market, May 2019

451 Perspective: Relax, sky is not falling for Hadoop, August 2019

451 Perspective: Open source, data platforms and the cloud - trouble in paradise, July 2019

NoSQL: A database market seeing changes as growth lies ahead, February 2019

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
23
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
6. Index of Companies
Amazon Web Services 14 HPE 1, 14, 18

Arcadia Data 1 IBM 14, 18

BlazingSQL 1 MapR 1, 14, 15, 18

BlueData 1, 18 Microsoft 14, 19

Cazena 14 OmniSci 1

Cloudera 1, 14, 15, 16 Oracle 14, 19

Cray 1, 18 Qubole 14, 19

Google 14, 17, 19 SAP 14, 20

Hortonworks 1, 14, 16 Ververica 1, 21

D a t a P l a t fo r m s M a r ket M a p 2 0 1 9
24
© C O P Y R I G H T 2 0 1 9 4 5 1 R E S E A R C H . A L L R I G H T S R E S E R V E D.
 About 451 Research
451 Research is a leading information technology research and advisory
company focusing on technology innovation and market disruption. More than
100 analysts and consultants provide essential insight to more than 1,000
client organizations globally through a combination of syndicated research and
data, advisory and go-to-market services, and live events. Founded in 2000
and headquartered in New York, 451 Research is a division of The 451 Group.
© 2019 451 Research, LLC and/or its Affiliates. All Rights Reserved. Reproduction and distribution of this publication, in whole
or in part, in any form without prior written permission is forbidden. The terms of use regarding distribution, both
internally and externally, shall be governed by the terms laid out in your Service Agreement with 451 Research and/or
its Affiliates. The information contained herein has been obtained from sources believed to be reliable. 451 Research
disclaims all warranties as to the accuracy, completeness or adequacy of such information. Although 451 Research
may discuss legal issues related to the information technology business, 451 Research does not provide legal advice
or services and their research should not be construed or used as such. 451 Research shall have no liability for errors,
omissions or inadequacies in the information contained herein or for interpretations thereof. The reader assumes sole
responsibility for the selection of these materials to achieve its intended results. The opinions expressed herein are
subject to change without notice.

N E W YO R K SAN FRANCISCO
Chrysler Building 505 Montgomery,
405 Lexington Avenue Suite 1052
9th Floor San Francisco, CA 94111
New York, NY 10174 P 212-505-3030
F 212-505-2630

LO N D O N B O S TO N
Paxton House 75-101 Federal Street
Ground floor 5th Floor
30 Artillery Lane Boston, MA 02110
London, E1 7LS, UK P 617-598-7200
P +44 (0) 203.929.5700 F 617-428-7537
F +44 (0) 207.657.4510

©2019 451 Research, LLC

W WW.451RE SEA RC H .C O M