Cloud Computing Today recently had the privilege of speaking with Amos Shaltiel, CEO and co-founder and Michael Elkin, COO and co-founder of DBS-H, an Israel-based company that specializes in continuous Big Data integration between relational and NoSQL-based data. Topics discussed included the core capabilities of its big data integration platform, typical customer use cases and the role of data enrichment.
Cloud Computing Today: What are the core capabilities of your continuous big data integration platform for integrating SQL data with NoSQL? Is the integration unidirectional or bidirectional? What NoSQL platforms do you support?
DBS-H: DBS-H develops innovative solutions for a continuous data integration between SQL and NoSQL databases. We believe that companies are going to adopt a hybrid model where relational databases such as Oracle, SQL Server, DB2 or MySQL will continue to serve customers alongside new NoSQL engines. The success of Big Data adoption will ultimately rise and fall on how easily information can be accessed by key players in organizations.
The DBS-H solution releases data bottlenecks associated with integrating Big Data with existing SQL data sources, making sure that everyone has access to the data they are looking for transparently and without the need to change existing systems.
Our vision is to make the data integration process simple, intuitive and fully transparent to the customer without a need to hire a highly skilled personnel for expensive maintenance of integration platforms.
Core capabilities of the DBS-H Big Data integration platform are:
1. Continuous data integration between SQL and NoSQL databases. Continuous integration represents a key factor of successful Big Data integration.
2. NoSQL data modeling and linkage to existing relational model. We call it a “playground” where customers can :
a. Link a relational data model to a non-relational structure.
b. Create new data design of NoSQL database
c. Explore “Auto Link” where engine automatically generates 2 options of NoSQL data model based on existing SQL ERD design.
3. Data enrichment – capability that allows to add to each block of data additional information that significantly enriches that data on the target
Currently, we focus on unidirectional integration and avoid some of the conflict resolution scenarios specific to bidirectional continuous data integration. The unidirectional path is from SQL to NoSQL and in the near future we will add the opposite direction of NoSQL to SQL integration. Today, we support Oracle and MongoDB databases and plan to add support for additional database engines such as SQL Server, DB2, MySQL, Couchbase, Cassandra and full integration with Hadoop. We aspire to be the default solution of choice when customers think about data integration across major industry data sources.
Cloud Computing Today: What are the most typical use cases for continuous data integration from SQL to NoSQL?
DBS-H: NoSQL engines offer high performance on relatively low cost and flexible schema model.
Typical use cases of continuous data integration from SQL to NoSQL are driven principally from major NoSQL use cases, such as:
- Customer 3600 view – creating and maintaining unified view of a customer from multiple operational systems. Ability to provide consistent customer experience regardless of the channel, capitalize upsell or cross-sell opportunities and deliver better customer service. NoSQL engines provide performance response time required in customer service, scalability and flexible data model. DBS-H solution is an enabler for a “Customer 3600 view” business case by doing transparent and continuous integration from existing SQL based data sources.
- User profile management – applications that manage user preferences, authentications and even financial transactions. NoSQL provides high performance, flexible schema model for user preferences, however financial transactions will be usually managed by SQL system. By using DBS-H continuous data integration financial transactions data is found transparently inside NoSQL engines.
- Catalog management – applications that manage catalog of products, financial assets, employee or customer data. Modern catalogs often contain user generated data from social networks. NoSQL engines provide excellent capabilities of flexible schema that can be changed on the fly. Catalogs usually aggregate data from different organizational data sources such as online systems, CRM or ERP. DBS-H solution enables transparent and continuous data integration from multiple existing SQL related data sources into new centralized catalog NoSQL based system.
Cloud Computing Today: Do you perform any data enrichment of SQL-data in the process of its integration with NoSQL? If so, what kind of data enrichment does your platform deliver? In the event that customers prefer to leave their data in its original state, without enrichment, can they opt out of the data enrichment process?
DBS-H: The DBS-H solution contains data enrichment capabilities during the data integration process. The main idea of “data enrichment” in our case is to provide a simple way for the customer to add logical information that enriches original data by:
- Adding data source identification information, such as: where and when this data has been generated and by whom. This can be used by auditing for example.
- Classifying data based on the source. This information can be very useful when customers what to control data access based on different roles and groups inside organization.
- Assessing data reliability as low, medium or high. This enrichment is useful for analytic platforms that can make different decisions based on source reliability level.
Customers can create enrichment metrics that can be added to every block of information that goes through the DBS-H integration pipeline. If no enrichment is required then the customer can opt out of the enrichment step.