01 is introduced

Today’s communications service providers (CSPS) need to be able to handle large volumes of complex data without slowing down or slowing down network response times and reliability. In the 5G era, the exponential growth in the number of devices and users has put forward new demands for business support services (BSS), which has become a particularly difficult task.

As you can see from the current reality, applications such as telecom network policy response, personalized quotes or fraudulent transaction prevention must be able to respond to data events in milliseconds to increase revenue or prevent losses.

To better meet these increasingly complex requirements, CSPS need to know how to best manage data in an increasingly crowded database environment, and new categories of these scenarios seem to emerge every year. The latest category is NewSQL, which offers telecom companies a unique advantage over NoSQL and SQL databases, especially in real-time data processing. Today’s databases traverse the entire life cycle of data, from fetch to execution, and must do so in 10 milliseconds or less. Looking around, only the NewSQL data platform can do this right now.

This paper explains the main differences between SQL, NoSQL and NewSQL databases, and explains why NewSQL database is the key for the telecom industry to adapt to the development of The Times, and how CSP can make full use of various database technologies for efficient operation and maintenance management of its network in the 5G era.

02 NewSQL origin

NewSQL is a term coined by Analyst Matt Aslett of the 451 Group to describe a new set of database features that inherit many of the capabilities of traditional SQL relational databases while providing some of the benefits of NoSQL technology.

The NewSQL system provides the best of both worlds for reality: ACID transaction consistency between a relational data model and a traditional database; Inherit the interactive convenience of SQL and the scalability and speed of NoSQL. Some systems offer greater consistency assurance than NoSQL solutions, and while “tunable” consistency is considered pseudo-consistency, it is not exactly ACID compliant.

Of course, there are differences between NewSQL solutions. SAP HANA can handle small amounts of transactional workloads without the benefits of a local cluster. NuoDB is a cluster-first SQL database focused on cloud deployment, but throughput is poor. MemSQL is useful for cluster analysis, but its tunable consistency is not strictly ACID transactions. Both NuoDB and MemSQL are computationally and story-separated, so they can have problems with data transfer and synchronization, especially around transactions.

The principle of the ACID

Most relational databases follow the ACID(atomicity, Consistency, Isolation, and persistence) principles, while most NoSQL databases follow the BASE(Basic Availability, Soft state, Final consistency) principles.

NewSQL databases, such as VoltDB, offer the scalability of NoSQL systems for online Transaction processing (OLTP) work, while adhering to the ACID guarantees of traditional database systems.

03 NewSQL and NoSQL in the Telecom Industry

Now that we’ve noticed the basic differences between SQL, NoSQL, and NewSQL, as well as their respective strengths and weaknesses. Let’s take a closer look at what features of NoSQL and NewSQL really matter to telecom operators and developers, and what problems they can solve with NoSQL.

  • What problems can I solve with NoSQL?
  • Where is NoSQL not appropriate?
  • How to take advantage of NoSQL and NewSQL?

We have no doubt that NoSQL databases are well suited for many work scenarios, but there are certain scenarios where NoSQL technology may not be the best solution to choose.

The next section compares NewSQL and NoSQL in four key metrics for data management in the telecommunications industry: scalability, availability, data consistency, and fast response.

3.1 Scalability

NoSQL

With 5G booming and the rapid growth of communication equipment, telecom companies need to upgrade and expand their existing data management methods.

NoSQL first came to attention when it was widely adopted in the Internet industry as a way to manage massive amounts of data like Google, Facebook, and Twitter. These platforms handle a large influx of unstructured data: Web searches, mobile devices, user status updates, information flows, etc.

In these scenarios, the most important consideration is scalability. Databases must scale massively and quickly. Relational schema and extension Traditional SQL databases cannot cope with massive data growth and processing, and the cost and efficiency of maintaining massive data and diversified query processing requests in traditional SQL databases are unacceptable.

The most important feature of NoSQL systems is the ability to extend applications on common hardware devices. NoSQL may be the right choice for requirements scenarios that require horizontal unlimited scaling, and NewSQL and NoSQL are not much different in extensibility.

However, NoSQL databases compromise almost everything else for scalability, which is problematic for telecom companies that rely solely on NoSQL.

NewSQL

Although NoSQL relational database systems offer scalability options, this is usually costly. The NewSQL system also addresses the scalability challenges of the system, while inheriting the transactional and SQL standards of traditional RDBMS.

In a typical scenario, a massively parallel SQL relational database in memory can scale linearly on general-purpose hardware. Like NoSQL solutions, NewSQL databases are cloud-friendly and can be scaled to meet the needs of applications with very large data scales. The system should be designed with no shared data blocks in the cluster to achieve low-latency read/write performance in the cloud environment.

NewSQL databases provide high availability, fault tolerance, and physical data redundancy, and run smoothly in scenarios such as telecommunications networks, so that telecom operators can cope with large amounts of incoming data. With the powerful NewSQL database, users can also build real-time transaction-oriented applications for real-time data flow processing scenarios.

3.2 availability

NoSQL

NoSQL systems are designed for the availability of CAP theory, which means that the database is always responsive even in the case of distributed partitioning.

But NoSQL systems are designed to prioritise availability and use final consistency rather than strong consistency (that is, always providing the latest and most correct snapshot of the data set), meaning that NoSQL systems can return less than the latest data in order to respond quickly.

Apache Cassandra is a follower of the idea of ultimate consistency, where quick response is more important than always returning the latest data, and indeed for many applications ultimate consistency is acceptable.

But in scenarios where transactions are based on exact data, such as telecoms companies taking steps to combat fraud, consistency is ultimately unacceptable.

  • Therefore, NoSQL solutions are not suitable for:
  • Decide whether or not to call a mobile user’s phone
  • Track (count) and allocate limited scarce resources
  • Transaction decision

NewSQL

NewSQL systems prioritize consistency over availability. The NewSQL system will return the same exact answer to all customers, enabling applications to make real-time decisions about call charges, airline seat assignments, and inventory without conflict.

3.3 consistency

NoSQL

As mentioned earlier, NoSQL systems are designed for scalability and availability, but at the expense of strong consistency. Therefore, NoSQL systems are not a good choice for scenarios that require strong consistency, such as billing and operational support scenarios, both of which are common in telecom operations.

The same goes for fraud. Telecom operators, especially in developing countries, are under enormous pressure to misuse SIM cards that can even be counted in containers, resulting in billions of dollars in annual losses. Solving the problem of wire fraud requires large scale, accurate, real-time computation of query caller accounts, which NoSQL databases cannot do.

NewSQL

While NewSQL systems are highly consistent and prioritize consistency over availability, NewSQL also supports multiple partitions, which is critical to telcos and their ability to provide uninterrupted service because it means the cluster can continue to work even if node-to-node communication fails.

3.4 Quick response to transactional scenarios

NoSQL

Quick response scenarios are common in modern scenarios. Although NoSQL solutions can generally improve data storage speed, they cannot provide strong and consistent transaction support on a large scale.

Transactional applications that need to be fast and scalable include:

  • Allow mobile phone connections while verifying the user’s balance
  • Trade at the best price
  • Show mobile ads to potentially thousands of users without breaking their advertising budget
  • Provide telecom service providers with strict SLAs to detect credit card theft before approving transactions

For such applications, NoSQL databases are usually not the best choice because processing events can occur millions of times per hour, per minute. Businesses in telecommunications, financial services, online gaming, advertising technology, and other industries need to be able to handle the concurrency and latency of event processing. Therefore, a scalable and strongly consistent transaction solution is a must.

NewSQL

The NewSQL system provides high scalability and consistency for modern applications. Even when processing massive amounts of data, multi-partition redundancy allows the system to scale linearly, helping applications respond accurately and quickly to customer requests.

04 Use NewSQL to build scalable modern applications

Both NoSQL and NewSQL provide data storage capabilities for building highly scalable applications. NoSQL data stores are ideal for high availability scenarios. The NewSQL system provides strong consistency and transactional interactivity, even in scenarios where consistency is preferred over availability in the event of a failure.

While nearly all NoSQL solutions offer scalability, VoltDB offers it and adds strong consistent transaction support.

VoltDB is incredibly responsive, consistent and scalable. Of all NewSQL solutions, VoltDB is the strongest and most flexible in the face of cluster failures, and our independent validation of availability has seen many customers operate stably in clusters in production environments for years.

VoltDB stands out in app scenarios that require strong consistency, including:

  • Dealing with increasingly complex policy and billing rules issues in telecom BSS and networks
  • Prevent fraudulent calls from occurring from post-call fraud detection
  • Instant offers to telecom customers to improve the subscriber experience and ARPU applies machine learning rules to detect and prevent iiot intrusions
  • Measure, monitor, and detect performance degradation to avoid unexpected outages

VoltDB is the most mature NewSQL system on the market and a cloud-native database. It supports ACID transactions in real-time data streams and is very well supported for local clustering and the Hadoop ecosystem. In addition, it also integrates the high throughput, low latency data processing features, is a very good system, data intensive applications in high performance, low latency, strong consistency requirements well in the scene, is widely used in strategy implementation, personalized recommendation, such as fraud or anomaly detection need real-time data flow in the application of response.

If you’re looking to integrate VoltDB into your tech stack, or want to talk to more of your friends

Please contact us by private message.

About the VoltDB

VoltDB supports strong ACID and real-time intelligent decision making apps to enable connected worlds. No other database product like VoltDB can fuel an app that demands a combination of low latency, massive scale, high concurrency and accuracy at the same time.

Founded by 2014 Turing Prize winner Dr Mike Stonebraker, VoltDB has redesigned relational databases to address today’s growing real-time manipulation and machine learning challenges. Dr. Stonebraker has been researching database technology for more than 40 years and has brought many innovations in fast data, streaming data and in-memory databases.

During VoltDB’s development, he realized the full potential of using in-memory transaction database technology to mine streaming data, not only to meet the latency and concurrency requirements of processing data, but also to provide real-time analysis and decision-making. VoltDB is a trusted name in the industry and has worked with leading organisations such as Nokia, Financial Times, Mitsubishi Electric, HPE, Barclays and Huawei.