NewSQL: The Bridge between SQL and NoSQL

Dulari Bhatt 23 Jun, 2022 • 4 min read

This article was published as a part of the Data Science Blogathon.

Introduction

NewSQL is more specific than NoSQL. Relational data and the SQL query language are the foundations of NewSQL systems. They want to address the NoSQL movement’s scalability, flexibility, and lack of focus difficulties. More consistency is provided with the new SQL. Within this category, however, there are a few differences. HANA was a robust business reporting platform that could also handle low transactional demand, making it perfect for SAP deployments. Hekaton adds enhanced memory processing to the more typical MS SQL Server. Both systems are currently non-clustering and are designed to quickly replace or supplement SQL deployments.

Suppose you have gigabytes to terabytes of data that require high-speed transactional access. You may have an input stream, such as sensors, mobile phones, or network access points, and you will require per-event transactions to compute real-time answers and analytics. Using the “ingest, analyze, decide” approach to tackle this issue, with analytics and choices determined per request rather than processing in batch processing. In this case, NewSQL systems, which combine the scalable architecture of NoSQL with greater consistency, are the best option.

Characteristics of NewSQL

1. SQL as the query language of choice
2. Transactions are supported with Atomicity, Consistency, Isolation, and Durability property
3. A concurrency control technique that isn’t locked (e.g., MVCC)
4. An architecture with significantly improved per-node performance
5. A shared-nothing, scale-out architecture

Completely new database platforms are the first type of NewSQL system. These are made to work in a cluster of shared-nothing nodes, with each node owning a portion of the data. These databases are frequently built up with architecture in mind, with features like concurrency management, flow control, and query processing.

Google Spanner, Clustrix, VoltDB, MemSQL, Pivotal’s GemFire XD, SAP HANA, NuoDB, TiDB, and Trafodion are examples of systems in this category.

Difference between SQL and NoSQL and NewSQL

There are some differences between SQL, NoSQL, and New SQL. The following figure depicts the difference very nicely.

 

Differences | SQL

                                                  Figure 1: SQL v/s NoSQL v/s NewSQL [1]

Parameters to compare

SQL

NoSQL

NewSQL

1. Scalability

Scalable Reads

Scalable reads and writes, horizontally scalable

Scalable reads and writes, horizontally scalable

2. Schema

Relational Schema / Table

Schema free

Supports both.

3. High Availability

Custom High Availability

Auto High Availability

Built-in High Availability

4. Storage

On-Disk +Cache

On-Disk +Cache

On-Disk +Cache

5. Cloud Support

Not Fully

Supported

Fully Supported

6. Query Complexity

Low

High

Very High

7. ACID-CAP BASE

ACID

CAP through BASE

ACID

8. OLTP

Not Fully Supported

Not Supported

Fully Supported

9. Security Concerns

Very Very High

Low

Low

10. Examples

Oracle, MsSQL, MySQL, etc.

MongoDB, Cassandra, Redis, etc.

CockroachDb, Google Spanner, VoltDb

The Most Effective NewSQL Databases

NewSQL is an intriguing technology that is becoming more widely available. The top NewSQL databases on the market today are adaptable and powerful solutions that have been meticulously designed to fit your company’s requirements.

VoltDB

VoltDB, founded in 2009, has lately risen to become one of the most prominent NewSQL providers in the booming in-memory database market. Python, Java, PHP, and C++ are among the languages supported. VoltDB is an open-source and commercial tool designed to operate with OLTP queries and includes ACID guarantees, sharding, and data replication. It employs deterministic concurrency management and a global serial order controller for sharded transactions that span many partitions.

ClustrixDB

ClustrixDB is a database that provides good peace of mind thanks to automated fault tolerance and a mechanism designed to preserve numerous data copies. Clustrix is compatible with MySQL and combines multi-version concurrency control (MVCC) with two-phase locking for input conflict resolution. It was founded in 2006 and was recently bought by MariaDB. Clustrix has various benefits over AWS and other popular cloud-based database servers, including sophisticated failover and fault-tolerance features and the capacity to expand and perform beyond 1TB. Additional features of the cluster-based cloud architecture include online schema modifications, linearly scalable, and online growth.

 NuoDB

NuoDB, founded in Cambridge, Massachusetts, is widely used by prominent FinTech companies. Dassault Systèmes, a French 3D design corporation, purchased it in November 2020 for the supply of scalable, multi-cloud distributed SQL database solutions. Dassault Système utilizes NuoDB to manage their worldwide database of online 3D component makers and clients, bringing together engineers and designers from all over the world.

CockroachDB

CockroachDB is an adaptable NewSQL database invented and built by Cockroach Labs that is utilized by a broad range of well-known businesses, including JPMorgan Chase, Cisco, and Best Buy. In addition, CockroachDB is a strong tool for organizational automation since it can simply link to cloud-based cluster solutions like Kubernetes. CoackroachDB, created by ex-Google personnel, is based on Spanner and uses an open-source distributed SQL database with horizontal scaling and a highly consistent transactional layer.

VMware Tanzu GemFire

This NewSQL database, previously known as Pivotal Gemfire XC, was bought by VMware in 2019 and renamed VMware Tanzu GemFire. Gemfire is a multi-tiered, geo-distributed, clustered cloud database with exceptional performance. Gemfire’s memory, horizontally scalable design distinguishes it from other NewSQL DBs by ensuring consistent low latency even during high usage.

NewSQL’s Advantages

  • It reduces application complexity by greater consistency and provides complete transactional support.
  • Elastic scalability that dynamically rebalances partitions
  • Instant event responses from real-time alerts
  • Resilient’shared-nothing’ architecture that provides high availability is just a few of Gemfire XC’s standout characteristics.
  • .NET and NodeJS support for different languages is frictionless.
  • New SQL uses common tools and SQL that is familiar.
  • It has more analytics, including SQL and extensions.
  • Many New SQL systems combine traditional data, such as RDBMS and query models, with NoSQL-style clustering.

NewSQL Limitations

  • No
    NewSQL systems are intended to be as general-purpose as SQL systems.
  • In-memory
    systems may not suit data sets larger than a few terabytes.
  • Only
    provides limited access to SQL systems’ extensive tools.

Conclusion

NewSQL is a subset of current RDBMS that aims to match the scalability of NoSQL systems for read-write workloads in online transaction processing (OLTP) while keeping the ACID (Atomicity, Consistency, Isolation, Durability) features of a traditional database system. The primary limitation of NoSQL is that it lacks the ACID feature.

  • This article explains NewSQL and the tools on which it can be performed. 
  • It also describes the significant differences between SQL, NoSQL, and NewSQL.
  • NewSQL is just like a bridge between SQL and NoSQL. NewSQL is a type of database solution that is gaining a lot of attention from the IT community.

Reference

[1] https://medium.com/rabiprasadpadhy/google-spanner-a-newsql-journey-or-beginning-of-the-end-of-the-nosql-era-3785be8e5c38

The media shown in this article is not owned by Analytics Vidhya and is used at the Author’s discretion. 

Dulari Bhatt 23 Jun 2022

Frequently Asked Questions

Lorem ipsum dolor sit amet, consectetur adipiscing elit,

Responses From Readers

Clear