March 2025 Summaries

Organizations choosing between custom-built and commercial data integration solutions must weigh control, cost, speed, scalability, maintenance, compliance, and reliance on internal expertise. Custom integrations offer precise ETL customization, tailored performance, and direct control over security and governance, but typically require substantial development investment, ongoing engineering support, and careful documentation to reduce knowledge-loss risks. Commercial off-the-shelf platforms provide pre-built connectors, vendor-managed updates and support, faster deployment, scalability, and often compliance certifications, although they can involve subscription costs, less flexibility for specialized requirements, and possible vendor lock-in. Building may suit organizations with highly specific processing needs and sufficient technical resources, while buying is generally positioned as faster and less operationally demanding for most businesses. CData Sync is presented as a hybrid option, combining more than 300 connectors, low-code or no-code ETL, SQL transformations, custom scripting, and cloud or self-hosted deployment to offer automation alongside extensibility.

CData Sync’s Q1 2025 release introduces Workspaces to help organizations organize ETL jobs, connections, and transformations into customizable, access-controlled environments for teams, projects, or production stages. The feature extends existing role-based permissions with workspace-specific policies, helping limit sensitive data access while supporting collaboration. The release also adds near-real-time change data capture for DB2 LUW and Informix, expands reverse ETL with Oracle as a source and new destinations including HubSpot, Zoho CRM, and Kintone, and improves write performance for Databricks and Redshift destinations. According to internal benchmarks, the performance updates can reduce destination write times by more than 2.5 times for large workloads while lowering memory consumption and compute requirements.

Salesforce Connect provides real-time access to external data within Salesforce but charges $48,000 annually for each external data source connection, making multi-source integration costly for growing organizations. The text presents CData as an alternative that allows multiple cloud, on-premises, ERP, database, and warehouse sources to be consolidated behind a single Salesforce Connect endpoint, reducing the need to purchase multiple licenses while retaining a required initial license. CData Connect AI is positioned as a low-code option for linking SaaS systems through a virtual data layer, while CData API Server is intended for databases and data warehouses that need to be exposed through a unified OData API. Beyond potential savings of tens of thousands of dollars annually, the approach is described as supporting easier scaling, live access without copying or migrating data, centralized IT management, improved governance, and reuse of the same connections for analytics and reporting.

Generative AI was a central focus at the Gartner Data & Analytics Summit 2025, where discussions emphasized that successful enterprise adoption depends on preparation rather than simply deploying new models. Organizations need integrated, timely access to both structured and unstructured data to avoid incomplete, biased, or inaccurate AI outputs. They must also modernize governance to strengthen data quality, provide explainability and auditability, and monitor bias and compliance risks, as inadequate governance could prevent many enterprises from realizing AI’s expected value. Open, composable data platforms were presented as a way to support flexible integration across cloud and on-premises systems while reducing vendor lock-in and enabling scalable deployments. Finally, companies need broader AI literacy, upskilled data teams, and specialized roles in governance and ethics so that technical and business users can manage AI systems and interpret their results effectively.

CData Connect AI is presented as a self-service platform for connecting live SaaS application data, including Salesforce, SAP, and NetSuite, to Databricks Lakehouse through Lakehouse Federation without building complex ETL pipelines. It supports more than 270 data sources, translates SaaS APIs into SQL-accessible endpoints, allows users to filter and prepare data before loading or querying it, and can enable real-time querying and updates. The service can be integrated with Databricks Unity Catalog for access controls, auditing, and security, while its setup involves configuring SaaS and Databricks connections and then querying the data as virtual tables. A customer example describes connecting filtered Salesforce opportunity data to Databricks in under 15 minutes, replacing weeks of ETL work. For broader, high-volume replication and ETL, ELT, or reverse-ETL requirements, CData recommends its separate CData Sync product.

CData has introduced free-forever community licenses for its Python Connectors and selected .NET Providers, alongside a new Developer Center offering learning resources for Python and .NET developers, with Java and SQL content planned. The initiatives are intended to reduce the effort required to connect applications to diverse APIs, SaaS platforms, databases, cloud services, and NoSQL sources by exposing them through standard SQL interfaces. CData drivers model API endpoints as relational tables, provide discoverable metadata, support SQL operations such as joins and CRUD commands, and optimize queries through source-side processing, pagination, and batching. The Developer Center includes conceptual overviews, setup guides, practical tutorials for tools such as Entity Framework, ASP.NET, SQLAlchemy, Pandas, and petl, plus troubleshooting materials, code samples, and datasets. Community license users can also access self-service support through the CData Community, where developers and company experts exchange guidance.

CData Connect AI is presented as a unified data connectivity platform designed to help finance and FP&A teams overcome fragmented systems, manual CSV exports, inconsistent schemas, and time-consuming reconciliation work. It connects live data from accounting platforms such as Xero, QuickBooks, Sage Intacct, and NetSuite with more than 270 applications and databases, including Salesforce, HubSpot, Workday, ADP, Stripe, and advertising platforms. Finance analysts can use it to consolidate multi-entity financial data in tools such as Power BI and Excel, reconcile transactions across operational systems, and combine sales, marketing, payroll, and financial information for real-time reporting, forecasting, ROI, customer acquisition cost, and headcount analysis. The platform also supports reusable query templates, scheduled report refreshes, and broader data connections such as Shopify and LinkedIn Talent Insights, with a free trial offered for prospective users.

Usage-based pricing is widely used in SaaS and cloud data tools because it offers low initial costs and charges customers according to consumption, but the model can become difficult to budget for as data volumes increase. Using Fivetran as an example, the text argues that monthly active row-based pricing can rise sharply from thousands to tens of thousands of dollars annually as organizations scale their ETL workloads, potentially shifting data growth from a business benefit to a budget concern. It presents CData Sync as an alternative that prices primarily by number of connections rather than row consumption, offering five connections and up to 100 million monthly rows for an $8,000 annual Standard plan, compared with Fivetran’s cited $8,550 cost for 1.5 million rows. CData further states that its higher plans permit unlimited replication volumes and cites a former Fivetran customer that reportedly reduced costs while growing to billions of replicated rows, positioning predictable annual pricing as a key advantage alongside flexible deployment and high-volume performance.

Financial services organizations must balance broad, efficient data access with strict security, privacy, and regulatory controls while managing rapidly growing volumes of diverse transaction, market, customer, and operational data. The discussion identifies three central architecture challenges: slow and costly implementation of centralized cloud data warehouses due to fragmented data sources and custom ingestion pipelines; the need to modernize cloud capabilities without abandoning sensitive on-premises systems or violating information-security requirements; and the difficulty of democratizing data access while maintaining governance over permissions, personally identifiable information, ownership, and quality. It argues that automated, standardized ingestion tools can accelerate centralization, secure connectivity can link on-premises and cloud environments, and enterprise semantic layers can provide governed self-service access through role-based controls and data masking. Examples involving NJM Insurance, a global insurer, and pension fund PGGM illustrate reported reductions in implementation time, costs, and data-access delays, suggesting that selective modernization and centralized governance can help institutions derive greater value from their data.