Data continues to grow in volume, variety, and velocity, resulting in new data management technologies. Recently, Deepti Srivastava, product manager for Cloud Spanner at Google Cloud, discussed how database requirements are evolving and how Google's Cloud Spanner is advancing a relational-NoSQL-convergence approach by giving customers the combined benefits of relational database structure with non-relational horizontal scale.
"The needs of customers have shifted dramatically from 10 years ago, when traditional databases could meet most business requirements. Over these years, relational and non-relational databases (and workloads) have started to converge," said Srivastava.
DBTA: How would you characterize the current state of database management?
Deepti Srivastava: Databases have been around for over 30 years and have been the backbone of most businesses for storing and analyzing business-critical data. Traditional databases evolved to become extremely feature-rich to support the increasing demands of applications on their datasets.
An interesting thing happened a few years ago. With the proliferation of the internet and global connectivity—which gave rise to the gig economy—the volume, variety, and velocity of data increased. As developers started to create ever-more ambitious applications, existing database technology couldn’t keep up. This is where you saw the advent of the NoSQL movement. Developers were willing to give up some features—like consistency and relational semantics—in favor of scalability and high availability.
However, it became apparent that the relational semantics and strong consistency offered by traditional databases was actually very useful and made for a simpler programming model. So application developers tried to work around these scale and consistency issues in the application layer, making applications brittle and hard to maintain. Fast forward a few years and with this complexity of management and proliferation of data, and now there is an increased demand for cloud-native, managed database offerings that reduce total cost of ownership and enable companies to modernize and build new applications quickly and easily.
Google experienced this arc internally as we grew and scaled. We came to the realization that a combination of relational semantics and scalability was going to be the best approach for building new applications and products. This is where the Spanner project was started and what we continue to invest in and improve upon today.
DBTA: What is having the biggest impact on data management now?
DS: Increased innovation in technology has created types of data and workloads that didn’t exist 10 years ago. As a result, today, companies are generating more data than ever before. As the nature of business has changed, the lines of relational and non-relational databases have also blurred. We’re seeing that operational and analytical use cases are converging because real-time business analytics are becoming crucial for businesses to differentiate themselves. Lastly, any given consumer is now interacting with a number of platforms daily—whether they be SaaS or consumer applications—creating a true global user base for a number of companies.
Despite this changing database landscape, there’s been a steady increase of net-new database usage in the cloud. This has driven database usage up and even created a need for next-generation databases to meet these business needs.
DBTA: How are organizations addressing this?
DS: As organizations continue to experience this massive increase in data, that they want to store and use to derive business insights using analytics and AI and machine learning, they are looking to cloud databases to help them scale, increase operational efficiency, and provide their users with the best customer experience possible. This is particularly evident in industries, such as retail, gaming, and financial services, all of which have experienced an influx of new uncharted customer data as their industries have been disrupted by new technology and an increasingly global consumer base.
DBTA: Spanner was added to the Google Cloud Platform in 2017. What customer issues does Spanner address?
DS: The needs of customers have shifted dramatically from 10 years ago, when traditional databases could meet most business requirements. Over these years, relational and non-relational databases (and workloads) have started to converge. We responded to this with the launch of Cloud Spanner—pioneering the convergence movement by giving customers the combined benefits of relational database structure, with non-relational horizontal scale. This allows for more time to focus on what is actually driving their businesses forward.
Cloud Spanner provides customers with a fully managed experience, with scale insurance—giving them peace of mind that as their business scales, they won’t need to re-platform. We also aim to lower the cognitive overload for customers with a focused portfolio that meets their needs, as opposed to taking a “database per workload” approach.
DBTA: What are its key differentiating characteristics?
DS: Cloud Spanner is an enterprise-grade, scalable relational database that offers scale insurance to support applications of all sizes. When companies are looking to adopt managed databases, they’re evaluating maintenance, total cost of ownership and time to market through an efficient developer experience. With Cloud Spanner, we’re addressing each of these needs. Our approach enables them to create an instance in just a few clicks and then scale it simply as the application’s needs change.
DBTA: How is the increase of global applications with demanding uptime requirements affecting data management?
DS: With the rise of SaaS software and transactional B2C applications, there’s an interconnectedness powered by data. This interconnectedness means that the majority of companies today require data management support for a truly global user base. This reality requires global consistency, reliability, and security—all in the internet age of scale.
Meeting this global demand can be complex for customers, but a managed cloud-native database service enables simplicity that can truly optimize database administration.
DBTA: What is the Spanner’s TrueTime system?
DS: Google’s TrueTime service is an accurate timestamp service powered by atomic clocks and GPS, which lets Cloud Spanner maintain consistent, synchronized copies of data in multiple locations. What this means, is TrueTime measures both clock time and an error margin for that timestamp. That uncertainty bound provides two values—the earliest timestamp it could possibly be and the latest possible timestamp. This enables a database like Spanner to provide strong consistency across regions, and even continents, allowing performant ACID transactions at any scale.
DBTA: In what scenarios is Spanner resonating most strongly with customers?
DS: When customers choose Google Cloud Spanner, they’re focused on increased velocity, improved reliability and managed data at scale. They also want a solution that enables them to modernize quickly and easily, wherever they are.
DBTA: Can you provide some examples of where Spanner is being used to solve customers’ vexing problems?
DS: A leader in mobile gaming, Bandai Namco Entertainment [BNE] deployed Cloud Spanner to manage their data at a global scale. BNE came to us as it was developing its Dragon Ball Legands mobile game, looking for a database that could scale with millions of players and had a reliable and low-latency network across regions. Once the company deployed Cloud Spanner, BNE was able to focus on improving the game functions rather than worrying about operating their databases.
Similarly, Cloud Spanner is also helping finserv customers manage the explosion of data that the industry is seeing. The UK-based fintech company, Blockchain.com is using Cloud Spanner to manage operational costs and improve the user experience. Blockchain.com’s users rely on its Blockchain Wallet to secure and use cryptocurrencies, but managing these massive amounts of datasets is not an easy task. With Google Cloud, the company was able to lower overall operational overhead costs, while taking away the worry of growing datasets. Take Azimut as well, which handles investment and asset management by sourcing information through data vendors. The company deployed Google Cloud to scale and address different kinds of data challenges, allowing Azimut customers to control their investments through the app, while Azimut manages the fund.
DBTA: How do you see the database market evolving over the next 5 years?
DS: Looking ahead, we’ll continue to see databases rooted in giving customers the fully differentiated value of cloud—no matter where they are. While we’re still in the early stages of cloud usage and growth, most organizations understand the benefits of cloud, like lower total cost of ownership and managed services. As more and more companies make the shift to the cloud, we’ll continue to see database vendors help businesses modernize their database strategies, quickly. Specifically, you’ll see Google Cloud lead in the cloud by offering a more efficient, integrated data platform.