About Us

banner
banner
banner
banner
bannerAbout Us

A Semi-Relational Causal Consistent Knowledge Base for Distributed Computing and Persistence with Homomorphic Quantum Secure Ledgers and Pods.

A zero-cost embedded database backed by LMDB, PostgresQL, and Prolog-AllegroSQL's imperative append-log interface. Exploiting SIMD: The Battle With accelerating point and range search queries as instances of the more general configurable combined compute and storage capabilities of memristor skews

  • banner

    EinsteinDB introduces novel configurations offering exclusive hybrid data structures that use both conventional CMOS processors/cache hierarchies and memristors for compute/storage!

    EinsteinDB is An embedded Datalog-AllegroSQL (locally immutable) in-memory database and Datalog query engine written in Rust, with Multi-Raft Haskell written consensus, MySQL and SQL parsing with Golang via MilevaDB, and TPC-C Benchmarking with Clojure Jepsen-Ready High-Throughput, Low-Latency Artifacts.

  • banner

    1000Txn/sec

    $0.020010Gbps for Read-Heavy workloads

  • banner

    99% Availability

    QoS on Ingress via Relativistic Sidecars

  • banner

    Under millisecond Reads

    Optimal Stochastic OLAP joins.

Our new data structures combine T-trees, B+- trees, and MemCAM to obtain a balance between search time and lifetime by exploiting a heterogeneous computing environment.

More About Us
banner

Our History

Inspired by the problems facing Automated Machine Learning on the Cloud and as a service, EinsteinDB sprung out as a persistence layer for Titus, CHAP, Mantis, and CI/CD pipelines with Jenkins, Jira, and Snowflake.

  • You create a database on minkowski spacelike- tuplespace, put some data in it
  • track changes, do queries and forget about it, or place it into the future heap: minkowski lightlike.
  • A persistent, embedded knowledge base with relativistic linearizable via VioletaBFT: A Multi-Raft invariant with HoneyBadger speculative BFT
  • makes it easy for you to grow to accommodate new kinds of data, for data to synchronize between devices, for multiple consumers to share data, and even for errors to be fixed.

Our Mission

A physical place, for our purposes, has an address. (It might have more than one: The introduction of fine-grained entities to represent data pushes us towards immutability: changes are increasingly changing an ‘arrow’ to point at one immutable entity or another, rather than re-describing a mutable entity. 100Gbps at $0.01000 our revolution.

  • designing CPU-efficient remote storage stacks
  • NVMe-over-RDMA
  • NVMe-over- TCP
  • In Kubernetes, Mesos, Docker and Ansible we trust

Who we are

Karl Whitford, Josh Leder, Joe Pollard, and Ligeia Mare are building practical asynchronous BFT protocols, in-memory databases for Computational Genomics, Quantum Encryption and Computing, Online Machine Learning and Multiplayer Online Game-Physics over TCP/IP

  • liveness guarantees without making any timing assumptions
  • We base our solution on a novel atomic broadcast protocol that achieves optimal asymptotic efficiency.
  • optimize set intersections and the associated data layout to be well-suited for SIMD parallelism
  • through a relativistic linearizable high-latency, high-overhead network reqs.
banner
bannerThe history of FIDel - From IBM to WHTCORPS

Similar to micro-services deployed in a cloud, FIDel (jobs) are deployed on the EinsteinDB platform. FIDel provides the APIs to manage the life cycle of jobs in both MilevaDB and EinsteinDB (like deploy, update, and terminate), manages the underlying resources by containerizing a common pool of servers, and, similar to a traditional micro-service cloud, allows jobs to discover and communicate with other jobs.

  1. 99.9999% availability

    EinsteinDB has been in production at WHTCORPS since 2018. It began at Netflix as project Mantis, grew into a book, and became a business in 2019. EinsteinDB processes trillions of events and peta-bytes of data every day.

    MilevaDB written in go for EinsteinDB written in rust is designed for business-critical database applications that require fast performance, high concurrency, and automatic scaling. You can scale up to millions of queries per second and 100 TB per database cluster with 15 low latency read replicas.

    MilevaDB decouples compute and storage resources, giving it 6 times faster than standard MySQL databases in high concurrency scenarios.

    banner
  2. 100Gbpsper second

    these architectures are well-known to cater for data rates of 100Gbps

    Each MilevaDB instance for EinsteinDB cluster supports to scale up to 100 TB and can be scaled out to up to 16 nodes. Each node can have up to 88 vCPUs.

    The relational nature of Prolog makes its co-habitation with relational database systems an attractive proposition. Not only databases can be viewed and used as external persistent storage devices that store large predicates that do not fit in memory, but it is also the case that Prolog is a natural choice when it comes to selecting an inference engine for database systems. T

    banner
  3. 2014March 24th

    From Stream Processing at Netflix to Bipartite Beyond-Relational Workloads at CloudKitchens.

    For Netflix to be successful, it has to be vigilant in supporting the tens of millions of connected devices that are used by the 40+ million members throughout 40+ countries. These members consume more than one billion hours of content every month and account for nearly a third of the downstream Internet traffic in North America during peak hours in 2014

    banner
  4. 2015November 10th

    Kubernetes

    Kubernetes launched with a great API and CLI that developers love. At Mesosphere, we saw its potential early and invested in bringing it to Mesos.

    banner
bannerPeople Love Us

Why Choose Us?

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna.

1

Data for All Your People

Dolor sit amet consectetur elit eiusmod tempor incidi dunt labore dolore magna aliqua enim.

2

A New Breed of AI

Dolor sit amet consectetur elit eiusmod tempor incidi dunt labore dolore magna aliqua enim.

3

Analytics Business

Dolor sit amet consectetur elit eiusmod tempor incidi dunt labore dolore magna aliqua enim.

banner
aboutMilevaDB's Interlocking Directorate helps manage MilevaDB on Kubernetes and automates operating tasks, which makes MilevaDB easier to deploy on our enterprise managed memristor-based Kubernetes offering.

EinsteinDB allows one to apply classical optimizations to multiway joins via stochastic foragers: TiDB works as the SQL layer and TiKV works as the Key-Value layer. MilevaDB provides EinsteinDB the SQL pushdown automata turning EinsteinDB into an Append-Only Hybrid HTAP.

The EinsteinDB storage engine manages user data and indexes. It provides transactional operations via MilevaDB on tables of records, hash and range index base mechanisms for storage, checkpointing, recovery and high-availability.  The EinsteinDB JIT VioletaBFT compiler takes an abstract tree representation of a T-SQL stored procedure, including the queries within it, plus table and index metadata and compiles the procedure into native code designed to execute against tables and indexes managed by the EinsteinDB's storage engine.  The EinsteinDB system is a relatively small component that provides integration with SQL Server resources and serves as a common library of additional functionality needed by compiled stored procedures.

1
about

EinsteinDB and MilevaDB offers better throughput at high loads, but slightly higher latency (still at hundred-microsecond granularity) than RocksDB, Vanilla PostgresQL, MySQL Inno, MonetDB, and AllegroGraph

Together with FIDel, MilevaDB(MySQL compatible auto load-balanced reverse proxy), and EinsteinDB, offer both High-density resistive ACID queries for Key-Value on BTree+. We virtualize and employ stealth FPGAs that speed up KOPS 10x using lookup tables (LUTs) to create specified parliaments.

2
about

EinsteinDB fits into a conceptual lineage that includes Freebase's graphd and 2005-onward Semantic Web stores

Globally distributed, ACID-compliant database that automatically handles replicas, sharding, and transaction processing, so you can quickly scale to meet any usage pattern and ensure success of your products.

3
about

With FIDel, MilevaDB, BerolinaSQL, and VioletaBFT, EinsteinDB excels in horizontal scalability and can easily grow to process up to 100+ terabytes of data, Millions of Rows per Second; Hundreds and Thousands of Interactive Transaction and Correctness Security Provisions on the fly..

Features Zero-copy API. Reads return references into the memory-mapped file. Cursors directly map to the same operations provided by LMDB, but in a typesafe manner. Nested transactions. Full integration with the borrow checker. Read references are checked to not outlive their transaction or overlap with a write in the same transaction. realtime read/write access to your Big Data.

Managed DBaaS
about

MilevaDB is the Optimized for fast random lookups and infrequent bulk inserts distributed SQL sitting on top of EinsteinDB. While EinsteinDB and MilevaDB are share-nothing databases. Prolog storage system for managing structured data that is designed to scale to a very large size have not been implemented at scale. petabytes of data across thousands of commodity servers are in need of a Larger-Than-Real-Life Solution: EinsteinDB provies larger-than-memory data set joins.

The MilevaDB API provides functions for creating and deleting tables and column families stored on EinsteinDB(Tree-SSD). It also provides functions for changing cluster, table, and column family metadata, such as access control rights.

5
about

using just-in-time (JIT) compilation to execute queries as native code instead of interpreting a plan.

The state-of- the-art in query compilation compelled us to fuse operators together (DeltaTables, Append-Logs, etc) in a query plan to minimize materialization overhead by passing tuples efficiently between operators.

6
about

EinsteinDB is built on several other pieces of WHTCORPS infra. EinsteinDB uses MilevaDB to store log and data files. A cluster typically operates in a shared pool of machines that run a wide variety of other distributed applications, and VioletaBFT processes often share the same EinsteinDB replica with processes from other SQL or NoSQL or Graph applications.

As DRAM becomes increasingly cost-effective, it enables EinsteinDB, MilevaDB to become memory-resident.

about
aboutTeam Members

Our Data Scientist

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna.

about

Merv Adrian

Data Management
about

Kirk Borne

Data Scientist
about

Carla Gentry

Analytical Solutions
about

Marie Curie

Data Scientist
aboutTestimonials

What Our Partners are saying

“The service control platform is the next-gen of traditional API management” Kong CTO and co-founder Marco Palladino

about