Run high performance SQL queries and batch jobs at a fraction of your cloud costs compared to alternatives. Querying large amounts of data is expensive and slow. Current solutions require huge amounts of compute and memory to be useful. With JitsuAI, an unlimited number of users can work concurrently on terabytes of data in their data lake using only gigabytes of memory. You can achieve linear price/performance with increasing memory and gain flexibility in managing SLAs while dramatically lowering your cloud costs.
Our products offer the highest level of cloud security by using a combination of machine-to-machine access that is integrated with your company's SSO/OAuth. Grants and privileges enable fine-grained Table/Column/Row level access controls. All of this is packaged as a complete PaaS offering within your own VPC with a fully managed service option.
Our management plane has an easy to use UX and also runs entirely within your own VPC to provide security to meet compliance & governance standards for banks, financial institutions, and healthcare industries. We can hand over the controls for you to manage your users while providing 24x7x365 enterprise support.
Simplify operations and spend less on databases and cloud infrastructure.
JitsuAI product use cases include:
· Ad hoc querying large amounts of data (an unlimited number of simultaneous users)
· Running batch jobs
· Accelerating your cloud BI infrastructure
Products Designed For You
~5x Lower Cloud Costs
Rapid Deployment, Easy to Use
With JitsuAI an unlimited number of business/data analysts can run high performance SQL queries directly on TBs/PBs data without the need for a data warehouse, while delivering extreme scalability at the lowest cloud TCO.
Cloud data warehouses are expensive proprietary silos. Use JitsuAI to simplify your operations and spend less on databases and cloud infrastructure.
Ease of Deployment & Use
At JitsuAI, we are focused on providing users with an ultimate end user experience. Our products are very easy to deploy and use. With just a few clicks users can get started and issue SQL directly on their data lakes.
Stay 100% Open
JitsuAI supports 100% open interfaces, helps avoid vendor lock-in, and as a result, creates more options for your analytics solutions. Its open-source components include SQL dialects (Spark & Presto SQL), indexing technology (Apache Iceberg, Microsoft Hyperspace), and file and table formats (Parquet, Avro, ORC).
The Power of Open Indexing
Apache Iceberg, Microsoft Hyperspace, and Apache CarbonData provide deep indexing, deltas, snapshots, schema evolution, meta data for point-in-time rollbacks, crash consistent recovery, and even coarse-grained transactions suitable for warehousing and analytics. Query optimizers can leverage these indexing standards to offer high performance and scalability.
There are many business reasons to choose open-source indexing technologies. Even as their capabilities continue to improve, they are already quite sophisticated and include features such as bloom filters. The indices are created during ETL. Presto and Spark have Iceberg plugins that allow the creation of Parquet, Avro, or ORC files along with rich meta-data and indexes. This keeps your ETL pipeline both open as well as efficient.
When Parquet files are created using an open indexing technology like Iceberg, the meta data and indexes that are created are richer and help with the query engine's Cost Based Optimization (CBO). This means faster, optimized and more efficient queries. Open indexing also levels the playing field for query processing. With a variety of indexing technologies available, and options for customers to pick based on their needs, it really comes down to—May the best query processing win! Why lock your meta data into a proprietary cloud DW?
Open source indexing should be an important consideration for businesses that want to adhere to open standards while cutting down costs.
Jumping Over the Scalability Wall
Ad hoc querying large amounts of data is expensive and slow. Current solutions require huge amounts of memory to scale with users and workloads. Processing a terabyte of data by using a terabyte of memory is expensive. In-memory style of compute is passe and not cost-effective for most analytics applications. New features such as "spill to disk," that are supported by most query engines, mitigate memory issues to some extent.
While pure in-memory systems can be faster, the reality is that despite their widespread use, resource utilization on the cloud is very low. Database workloads hover around 30% CPU utilization. Compressions and rarefactions of in-memory working sets when executing SQL result in “out of memory" exceptions and crash the system. Businesses spend more time trying to size their workloads to fit into memory than working on their business logic. Sizing the system for the worst case workload results in low resource utilization for the most part, and as a result, increases cloud costs. The price for compute on cloud is the same per second/hour/minute regardless of whether you utilize 30% or 100% of the CPU. There is the hidden cost of developer/analyst productivity where skilled technicians are required to manage such systems for businesses in geographies where such skills are hard to find, aside from being secondary or tertiary to their core business.