Key Concepts and Architecture

Applies to: Kyvos Enterprise  Kyvos Cloud (SaaS on AWS) Kyvos AWS Marketplace

Kyvos Azure Marketplace   Kyvos GCP Marketplace Kyvos Single Node Installation (Kyvos SNI)

Kyvos installs directly on clusters located on the cloud and on-premise data lakes. 

Once deployed, Kyvos uses the computing capacity of modern data platforms to process multi-dimensional OLAP semantic models on data at a massive scale. These semantic models pre-aggregate data across multiple dimensions and measures and are stored in a distributed manner across the data infrastructure. As all the heavy lifting is done in advance, these semantic models provide instant responses to queries, enabling quick data analysis. In addition, Kyvos maintains the native data security layer through the semantic models at the storage layer itself, ensuring that users only see the data to which they have access.

The following figure illustrates a high-level Kyvos architecture.

Open

Data platforms

Kyvos offers native support for Amazon Web Services (AWS), Google Cloud, Microsoft Azure, Cloudera, Hortonworks, MapR, and Apache Hadoop. It also supports all cloud data warehouses, such as Snowflake, Amazon Redshift, and Google BigQuery, and the latest releases of Cloudera, MapR, and Apache Hadoop.

As Kyvos processes and serves queries directly on your data platform, there is no need for any data movement. This saves you from the trouble of sending high-volume data over the network and ensures that your data is safe. Further, the Kyvos semantic models are stored on the data platform, so there are no additional infrastructure requirements. The Kyvos BI server and query engines can be deployed directly on your existing cloud or on-premise data platform.

BI acceleration layer 

Kyvos processes a BI acceleration layer on the cloud and on-premise data lakes, bridging the gap between your data and the business intelligence and analytics tools. This layer provides a consistent semantic model for business users, making it easy for them to visualize massive data. It allows them to see the dimensions and measures available to them and drag and drop them into their visualizations intuitively using the BI tool of their choice. 

The next layer of the Kyvos architecture consists of the BI server cluster, which enables the aggregation of data at a massive scale. The BI servers are designed with active-active load balancing for scalable deployments to thousands of users across the enterprise. The built-in load balancer intelligently routes queries to the correct server, maximizing performance and capacity utilization. Multiple BI servers ensure high availability with minimal downtime. In the case of a BI server failure, another one is automatically promoted for use.
When the user fires a query, the BI tools connect to the Kyvos BI server using standard connectors. The BI server parses the query and routes it to the query engines, which execute it and return the results to the BI tool. The Kyvos query engine cluster is highly elastic. The number of active query engines can be easily increased or decreased to deal with varying loads, accommodate more users, or further reduce response times. Kyvos query engines are optimized to return most queries in less than a second.

Besides this, Kyvos has an intelligent, multi-level caching mechanism that ensures high performance based on query patterns and usage. It stores segments of the most frequently queried semantic models and caches the results of the most frequent queries in high-performing storage.

Business intelligence and analytics tools

The topmost layer of the Kyvos architecture consists of BI and analytics tools that connect to Kyvos semantic models using standard access mechanisms such as SQL and MDX, enabling users to access massive data instantly and interactively using their choice of BI tools. Kyvos supports all major BI tools, including Business Objects, Cognos, Excel, MicroStrategy, Power BI, Qlik, Spotfire, and Tableau. Kyvos semantic models can also be accessed from data science engines like R and Python to discover data patterns. It supports REST and JAVA APIs that enable integration with custom enterprise applications.

Kyvos also has a native visualization engine with an intuitive drag-and-drop interface for self-service analysis.

Elasticity

Kyvos is architected to optimize resource utilization, deal with peak loads, and deliver cost-effective BI on the cloud. The solution can quickly scale up and down without disruption to deliver consistent performance and ensure optimal utilization of resources. As more data gets added to the BI environment, Kyvos scales out transparently to process semantic models on that data. Similarly, querying capacity can be scaled up or down depending on the expected load. This feature also allows it to support thousands of concurrent users without any impact on performance.

Security

Kyvos offers high-class, enterprise-level security through its built-in security model and support for standard security frameworks and protocols. It enables granular access control with column and row-level security built into the system. Additionally, it integrates with enterprise security systems like Knox, Ranger, and Sentry, as well as custom security frameworks. Kyvos supports Kerberos authentication and LDAP for enterprise user management and integrates with enterprise single sign-on tools such as SiteMinder and Okta.