Overview
Run Analytics at the Granular Level
As projects get started, they’re usually designed for a specific use case. This approach of a single use case and one data product at a time won’t scale. What’s needed is the connected multi-cloud data platform for enterprise analytics that can scale to answer any question at any time against any data by any user.
A platform with multidimensional scalability increases capabilities across eight core dimensions simultaneously without adversely impacting other domains. Scaling at this level delivers the advanced capabilities enterprises of the future need by empowering them to run millions of models on trillions of interactions, every second of every day. This type of scalability also lets companies run analytics at a granular customer, service, and monetary level.
Overcome Data Analytic Complexity
To handle a significant increase in concurrent queries, these technologies require users to:
-
Reduce the complexity of their queries
-
Implement separate platforms
-
Add a new cloud cluster
-
Limit users
-
Highly tune specific queries
These actions limit value, add complexity, reduce agility, and increase costs. They also limit future agility, restrict the ability to ask cross-functional business questions, and limit user flexibility to ask any question at any time. That’s why the connected multi-cloud data platform for enterprise analytics must offer the ability to hyperscale.
5 Requirements for Scalability in a Cloud-First World
To scale in a cloud-first world, companies need a modern cloud data analytics architecture that supports these five elements:
-
Separation of compute and storage with elastic scaling.
This modern capability supports the demands of both data and users without excess, unused capacity. The elasticity enables users to dynamically scale up/down and start/stop resources without IT intervention.
-
Integration with first-party cloud services.
Tightly integrating with services from Amazon Web Services, Microsoft Azure, and Google Cloud can accelerate data analytics ecosystem solution deployment.
-
Ingestion of modern data sources.
Supporting multiple data types in a single system helps integrate data, eliminate data and process redundancy, and support advanced analytics.
-
Integrated data management and scalable analytics.
A data analytics platform should unify analytics and data management to enable data exploration, modeling, and scoring at scale in a single, easy-to-use environment and automate data management functions.
-
Dynamic resource allocation and workload management.
A platform must have the ability to manage system resources and user workloads dynamically.
8 Core Dimensions of Multidimensional Scalability
Many vendors define a scalable platform as one that runs multiple queries and meets future data growth needs. That definition doesn’t go far enough. Scalability must also consider important factors such as latency, performance, reliability, availability, and total cost of ownership.
Scalability requires eight essential domains:
-
Data Volume
Efficiently store and process petabytes of data natively and in object storage. This makes it easy to access all the data needed to drive deep insights.
-
Query Concurrency
Simultaneously process large volumes of queries to get more work done faster. Optimize and balance numerous complex, resource-intensive queries while maintaining service level agreements (SLAs).
-
Query Complexity
Support complex, high-value queries, including multi-join queries. Ask questions that span different business functions and uncover new insights.
-
Schema Sophistication
Extensible and flexible data schemas support and enable all business requirements. This flexibility supports any schema—normalized, semi-structured schemas, or no schema.
-
Query Data Volume
This is the volume of data that can be processed quickly and efficiently by a single query, without manual intervention. With the right platform, business users don’t have to limit queries because all data can be considered.
-
Query Response Time
Deliver fast and consistent response times to comply with strict SLAs.
-
Data Latency
Load and update data in near real-time while also supporting query workloads. Users stay in sync with business processes and can respond to analytic needs against current data.
-
Mixed Workload
Support multiple applications/users with very different SLAs in a single environment to simplify workload management, help guarantee SLAs, and maximize resources.
These abilities to scale let companies ask sophisticated and new questions of all their data to achieve their goals. For example, multidimensional scalability helps companies like Royal Bank of Canada use modern data analytics. What can it do for you?