Relational databases—built well before Web 3.0—are bottlenecking cloud-native development and demand for real-time data agility. Can NoSQL solutions help?
With the global datasphere skyrocketing 15-fold in the last decade, relational databases—although still useful—can no longer cope with today’s applications, user needs, and scale out volume. One way to fill the gap is to rely on NoSQL, which provides a mechanism for storage and retrieval of data that is modeled in means other than the tabular relations used in relational databases.
However, implementing NoSQL systems into organizations is not easy. What incentives do they offer to organizations to invest time and money to adopt them? What considerations determine whether NoSQL is the right solution for CIOs in APAC?
DigiconAsia find out the answers from Deb Dutta, General Manager (Asia Pacific and Japan), DataStax.
DigiconAsia: What market dynamics are rendering traditional relational databases less compelling and creating a need for cloud-native database solutions?
Deb Dutta (DD): Relational databases have been a staple of modern computing since their conception in 1970. They were built during a time when data was mostly structured and was clearly defined by their relationships.
Today, we know that data is expanding in volume and structural complexity. That is where NoSQL and especially the open source Apache Cassandra, becomes relevant: they are built for planetary scale and real time applications that people now take for granted in our daily lives and businesses.
Market dynamics making NoSQL more compelling include:
- Built for Web 3.0: Blockchain, cryptocurrencies, smart contracts and real-time payments are require extremely scalable and powerful databases to power them.
- Unrivaled scale: Our digital world runs in real time. Today there are more real time applications powering customer experiences and driving new revenue models than ever before. At this scale, relational databases are being taxed to the limits while applications built in NoSQL are just starting to prove their good performance and unrivaled scale.
- Manageable total cost of ownership: NoSQL cloud databases offer pay-as-you-grow pricing. A developer can sign up for a NoSQL cloud database instance and build the coolest, high-growth applications quickly. As application adoption increases, or decreases, the cloud-based architecture expands or shrinks as needed so no one has to over pay for unused capacity.
- Freedom from rigid schemas: NoSQL offers a more flexible data model and freedom from rigid schemas. You may also see significantly improved performance and the ability to horizontally scale out the data layer.
DigiconAsia: How do CIOs determine if they should switch to NoSQL to modernize legacy applications? What are the risks and benefits?
DD: Unlike traditional SQL-based relational databases, NoSQL databases can store and process data in real-time.
CIOs and their developers wanting to build modern, real-time applications that have complex, constantly changing data sets and require flexible data models prefer NoSQL to bring their real time applications to market faster.
These applications engage customers and have the potential to drive new revenue models, so time to market is critical. Developers like the agile features that allow them to develop faster.
DigiconAsia: Can relational database systems also be modernized to regain relevance and competitiveness?
DD: There is debate, but we prefer to stick with our opinion that NoSQL databases like Cassandra, originally developed inside Facebook in 2008, were known from the ability to scale quickly and the capability to access huge amounts of data with very little latency.
For enterprises with a global presence—or those aspiring to such—Cassandra handles geographically distributed data very well and allows for cost-effective geographic expansion with commodity storage.
Because of its distributed nature, NoSQL is one of the most reliable databases for mission-critical applications. These are all capabilities that relational and SQL databases—the old guard—do not have in their DNA, and one reason Netflix abandoned its legacy database over a decade ago to rely instead on Cassandra. It just could not trust its business to technology with limited scalability and reliability.
DigiconAsia: How can organizations ‘try out’ NoSQL alongside existing database infrastructures before making the decision to implement this technology? Can it exist seamlessly alongside existing relational databases?
DD: Attempting to be a digital-first organization without a modern technology architecture to support it is impossible. What is possible is for such organizations to try out NoSQL with a few pilot projects and leverage pay-as-you-go cloud database technology.
There is zero infrastructure to manage; zero cloud vendor lock-in and modern APIs can get the project up and running quickly. NoSQL can sit alongside a relational database but handles operational data so it is like apples and oranges and not really comparable.
DigiconAsia: What NoSQL solutions are available out there, and how does each vendor differentiate its offerings? How can CIOs navigate through the differentiated offerings to find one that fits into their own organization’s appetite for change?
DD: There are a few vendors in the market and many of them are saying the same thing—making it difficult for CIOs to know which vendors to evaluate.
This is where organizations should seek a solution that creates value for enterprises and developers and differentiates itself, for example, by offering a ‘data stack’ that services operational data at rest via Cassandra, and data in motion through CDC and event streaming to provide real time business and customer-facing insights.
This example of ‘dual service’ approach offers organizations more value and utility when seeking a NoSQL solution that differentiates itself in the market.
DigiconAsia thanks Deb for his NoSQL insights.
