1. Solution Value
Acting on data the moment it is generated is the new business imperative for satisfying customers, succeeding with cross-selling and upselling efforts, and capitalizing on fleeting opportunities. Examples include delivering personal recommendations while a customer makes a purchase, mitigating the execution of stock trades, blocking a fraudulent payment before approval, and assisting patients when they require immediate attention.
To meet the demand to process large volumes of data in real time, businesses have a choice. Deploy individual capabilities to collect, manage, and process data or invest in a streaming data platform that covers these activities from end to end.
Integrated platforms provide extremely fast processing and analytics on streaming data sources like website clicks, sensor data, Internet of Things data, and continuously generated machine data. They solve the challenge of transforming largely unstructured and semi-structured data into a consumable format for analyzing and acting on that data. The best ones do so in real time, a term typically referring to completing a task or function within the time specified for doing so in a business SLA.
Hazelcast’s real-time stream processing platform utilizes a fast data store and real-time stream processing capabilities that are integrated with MLOps to provide in-the-moment action for fleeting business opportunities. Its low latency and extreme responsiveness allow organizations to achieve business objectives and monetize data that would otherwise not be possible.
2. Urgency and Risk
Real-time action on freshly created data is becoming a business imperative for organizations across verticals. The ability to process data as it is generated for instant analysis and action is a competitive advantage vital for attracting and retaining customers. Traditional batch processing is too slow for modern use cases in adtech, e-commerce, financial services, healthcare, supply chain, and logistics.
Delaying implementation of a streaming data platform increases organizational risks for churn, operational inefficiencies, poor agility, and missed business opportunities. Meanwhile, deploying individual components from multiple vendors (such as a fast data store or MLOps) increases complexity and fragmentation within the data architecture and can impact scalability.
Risk
The primary risk associated with deploying a streaming data platform is being unable to leverage the results for organizational reasons. For example, getting a real-time alert about a potential fraud case is one thing, but somebody must be in a position with sufficient authority and expertise to act upon it.
To prevent this outcome, a transition to real-time data processing requires a shift in mindset. Organizations must fully understand business processes and structures to merge and enrich the continuous flow of real-time data with relevant data from other systems. Scalability is another concern; the greater the scale of applications, the greater the amounts of quality data that must be processed for better insights. Deployments that don’t scale well won’t deliver real-time capabilities over time.
3. Benefits
The primary benefits of implementing Hazelcast include the ability to quickly build business applications to ingest huge flows of high-velocity, continuous data and enrich it with reference and other data sources to run rules engines and operationalize pre-built models that act on the data in-flow. Specifically, Hazelcast supports advantages such as:
- In-the-moment action: The platform enables action to data-driven events before they are over, such as responding to a customer’s click-screen patterns for real-time ads. Adtech is projected to expand the internet advertising industry to $1 trillion by 2030.
- Ease of implementation: Hazelcast’s platform includes all components for data enrichment, stream processing, and data preparation instead of requiring organizations to build these components piecemeal.
- Increased productivity: Organizations can achieve core mission objectives better and springboard to additional use cases from initial ones, like fraud detection. Consumers lost nearly $6 billion to fraud in 2021, and that doesn’t include remediation costs.
- Faster go-to-market: Organizations can respond to business changes by deploying solutions quickly. Hazelcast requires less development than other streaming data solutions that don’t have built-in data enrichment capabilities and the ability to operationalize machine learning models. It also provides multilanguage support so developers can simply design, build, and deploy applications. This lets them discover “what if” scenarios that uncover priceless revelations
4. Best Practices
Prudent organizations will thoroughly assess the requirements for their streaming data platforms. The best approaches involve considering their real-time business needs. Firstly, business owners should establish acceptable SLAs for delivering results to end users or other processes that need to act on the data automatically. They must also identify the necessary business requirements.
Next, organizations must define technological requirements. Considerations include which data to use and the diversity of source data (data complexity and format such as relational, object-oriented, hierarchical, or flat files), data models, and data velocity. Users must also address standards of quality for transforming and cleansing data, as well as the particulars of data integration/aggregation. All efforts should be made to minimize complexity when defining data integration/merging. Finally, they should ascertain which reference data sources are required for suitable enrichment of streaming data sources. Ideally, avoid data sparseness and have all necessary information for timely, informed action.
5. Organizational Impact
The organizational impact of deploying Hazelcast is considerable. Successful implementations will increase expectations about the time-to-value for data-driven initiatives, which will become drastically lower for achieving business outcomes. Thus, data culture and the worth of data will expand throughout the organization. Additionally, implementing this solution will spur stakeholders to expand their use cases and adopt others because of the newfound sense of empowerment true real-time processing provides.
People Impact
The use of Hazelcast will impact employees in several distinct ways. In terms of training and skills acquisition, technical users (including solutions architects, administrators, and IT teams) must become versed in streaming architecture and operation. Specifically, they need to become proficient in the Kappa architecture. Training could take as long as three months, depending on employees’ familiarity with this architecture. Three areas are impacted:
- Data leaders/business users: Rapid deployments enable internal personnel and external customers to consume real-time responsive applications while leveraging SQL data structures for broad access. Operationalized machine learning can automate time-sensitive applications.
- Architects/development: The integrated platform’s stream processing, fast data store, and ML components simplify architecture design while offloading integration, maintenance, and management burdens. Multilanguage support and out-of-the box connectors offer gateways to widely used data sources across on-premises and cloud deployments.
- Operations: Ops personnel can future proof business systems and leverage built-in connectors to integrate with modern, cloud-native applications. The Management Center Console (an enterprise feature for paying customers) monitors cluster performance, observes data about clusters, clients, and data structures, executes SQL queries on the cluster, and performs administrative tasks. The platform provides high-availability features and zero-downtime upgrades that improve TCO.
The budgetary impact of adopting Hazelcast will be largely positive, especially when compared to piecing together the individual components of a real-time solution. For example, Hazelcast accounts for data preparation, stream processing, machine learning inference, and low-latency storage in one platform, which is more cost-effective than purchasing and maintaining these elements individually. However, organizations may incur expenses related to scaling out for their throughput needs, which vary according to use case.
Investment Outlook
Hazelcast is available as a free, open-source solution and as a commercial offering (Hazelcast Platform Enterprise Edition). The open-source edition only provides community support, while the enterprise edition adds professional technical support and enterprise-specific features. The licensing for Hazelcast Platform Enterprise Edition is based on an annual subscription for each contract.
Customers pay according to the number of Hazelcast server nodes, called Members and Lite Members. One Hazelcast node is a Java Virtual Machine (JVM) operating an instance of Hazelcast.
Although there aren’t charges according to CPUs or CPU cores, there is a three-node minimum for a cluster. Other considerations impacting the number of JVMs in use include running multiple instances of Hazelcast on a single JVM, which counts as just one node. Hazelcast instances can run independently from the command line server or embedded within applications. Hot standby, warm standby, and production nodes are included in pricing; these nodes involve a JVM running Hazelcast’s software.
The platform is also available as a cloud managed service, called Hazelcast Viridian Cloud Managed Services, which appeals to businesses taking a hands-off approach to managing infrastructure resources.
6. Solution Timeline
Depending on the complexity of the use case, organizations can expect to go from proof of concept (POC) to production in approximately six months. This quick implementation is attributable to Hazelcast’s ease of use, multiple connectors, and simplified architecture.
Plan, Test, Deploy
To successfully deploy in six months—quick for streaming data applications—organizations must ensure they have all their business and technological requirements in place before beginning. Hazelcast offers options for painless, managed service deployments via cloud-native technologies. The vendor also provides expert services personnel to help organizations with the planning and testing stages.
Plan: This stage involves creating a POC, compiling the use case requirements, identifying data sources, and deciding which streaming platforms to use. Organizations should also determine their data preparation needs and ascertain which reference data is necessary for enriching the streaming data.
Test: The testing stage entails starting applications at a modest scale, focusing on monitoring capabilities, and iterating rapidly based on what’s learned. Hazelcast offers expert services (and has similar relationships with partners) to assist customers with workshops, diagrams, and delivery of services for core and associated applications.
Deploy: The deployment stage involves implementing Hazelcast into production and devising feedback mechanisms for assessing its outcome and ROI. Hazelcast’s cloud-managed services—Veridian Serverless and Veridian Dedicated—let organizations deploy in environments without worrying about the operational impact of their applications.
Future Considerations
Customers and prospects can anticipate Hazelcast simplifying its user experience so any developer, solution architect, or enterprise architect can deploy the platform for their application. Later releases will feature zero-code connectors and more drag-and-drop, point-and-click functionality to broaden adoption.
7. Analyst’s Take
The shift to real time is only increasing across industries and use cases as businesses leverage the ability to act in the moment for new revenue, risk reduction, and other efficiency-focused opportunities. Customer expectations are driving this realization. Hazelcast’s real-time stream processing platform fulfills this growing need by delivering low millisecond responses for inter-machine activity and responses with minimal latency for improved human decision-making and automated action.
8. Report Methodology
This GigaOm CXO Decision Brief analyzes a specific technology and related solution to provide executive decision-makers with the information they need to drive successful IT strategies that align with the business. The report is focused on large impact zones that are often overlooked in technical research, yielding enhanced insight and mitigating risk. We work closely with vendors to identify the value and benefits of specific solutions, and to lay out best practices that enable organizations to drive a successful decision process.
9. About GigaOm
GigaOm provides technical, operational, and business advice for IT’s strategic digital enterprise and business initiatives. Enterprise business leaders, CIOs, and technology organizations partner with GigaOm for practical, actionable, strategic, and visionary advice for modernizing and transforming their business. GigaOm’s advice empowers enterprises to successfully compete in an increasingly complicated business atmosphere that requires a solid understanding of constantly changing customer demands.
GigaOm works directly with enterprises both inside and outside of the IT organization to apply proven research and methodologies designed to avoid pitfalls and roadblocks while balancing risk and innovation. Research methodologies include but are not limited to adoption and benchmarking surveys, use cases, interviews, ROI/TCO, market landscapes, strategic trends, and technical benchmarks. Our analysts possess 20+ years of experience advising a spectrum of clients from early adopters to mainstream enterprises.
GigaOm’s perspective is that of the unbiased enterprise practitioner. Through this perspective, GigaOm connects with engaged and loyal subscribers on a deep and meaningful level.
10. Copyright
© Knowingly, Inc. 2023 "CxO Decision Brief: Streaming Data Platforms" is a trademark of Knowingly, Inc. For permission to reproduce this report, please contact sales@gigaom.com.