MMS • RSS
Listen to this story
Last June, DataStax, a real-time data company, raised USD 115 million in funding led by Goldman Sachs Asset Management. Founded in 2010, the Santa Clara-based company serves as the open, multi-cloud stack for modern data apps. With a valuation of USD 1.6 billion and a 700-strong employee base, the firm helps enterprises mobilise real-time data to build the smart, high-scale applications required to become data-driven businesses.
Deb Dutta, general manager of DataStax for the Asia Pacific & Japan region, had a freewheeling interaction with Analytics India Magazine, where he spoke at length about DataStax’s plans, its flagship products, how it deploys AI and ML and the shift towards NoSQL database management. Here are the excerpts:
AIM: DataStax has recently raised USD 115 million in funding. How do you plan to put it to use?
Deb Dutta: DataStax will leverage the funding round led by Goldman Sachs to accelerate the global expansion and development of its Astra DB multi-cloud database and its Astra Streaming streaming service, which are part of the company’s open data stack for building and running real-time applications on any cloud, at massive scale, anywhere in the world, with zero downtime.
AIM: Astra DB and Astra Streaming are two flagship products of DataStax. How are they different from their competitors?
Deb Dutta: Astra DB is a multi-cloud database-as-a-service (DBaaS) platform built on the world’s most scalable database – Apache Cassandra. It is designed to simplify building microservices-powered high-growth applications with the ability to reduce deployment time to mere minutes. Among the key features that will encourage developers to deploy Astra DB is the cost savings – which is thanks to its serverless pay-as-you-go pricing model and ability to scale up and down dynamically. This effectively reduces operational overheads and helps developers work efficiently while creating more time and space for innovation. From a security perspective, all data in Astra DB is encrypted both at rest and in motion, complemented by single sign-on and role-based access control.
Astra Streaming, on the other hand, is an advanced, massively scalable streaming service and cloud messaging platform built on top of the open source, distributed messaging and streaming platform Apache Pulsar. DataStax recently announced Astra Streaming with a new Apache Kafka connector and built-in support for RabbitMQ and JMS that can turn all enterprise data-in-motion into real-time streaming data and help give Kafka applications a “makeover”. With Astra Streaming, we aim to provide developers with a complete toolkit for data streaming backed by a compatible operational platform that provides zero-ops, infinite scale, and multi-cloud support as built-in features.
With Astra Streaming integrated into Astra DB, DataStax delivers an open stack that unifies all aspects of real-time data.
AIM: How do you deploy AI/ML to mobilise real-time data?
Deb Dutta: We use Apache Cassandra for executing large datasets. Thanks to its distributed structure, Cassandra is the only database where you can scale your data linearly and minimise cloud cost, making it the most elastic database in the market.
Using data that originates in applications and is streamed, stored, and immediately analysed to initiate action translates into valuable interactions with customers — and more customers. This can create a customer-facing virtuous cycle if you can re-deploy more types of data to improve existing experiences or offer new ones.
AIM: Apache Cassandra is the driving force behind DataStax’s services. What explains its popularity?
Deb Dutta: Apache Cassandra is widely considered to be the world’s most scalable database with the ability to power today’s modern applications. It’s a peer-to-peer platform, essentially infinitely scalable. It enables organisations to process large volumes of fast-moving data in zero downtime. Recent advancements, such as the inclusion of Stargate, an open-source data API layer, allow Cassandra to support multiple development tools, frameworks and data models, opening it to a wider developer community.
Interestingly, Cassandra was always known for its performance, versatility, and scale, but not for its ease of deployment or low cost. Astra DB addressed these issues by eliminating the complexities of deployment, software patching, and updating. Now, enterprises can deploy applications on Astra DB in minutes rather than weeks. Astra DB also made Cassandra ready for microservices prime time and simplified operations by fusing with Kubernetes as the centralised control plane while potentially reducing ownership costs by 75% over a three-year span.
Here are 5 reasons why it’s a great tool for ML and DataStax’s approach:
- Unparalleled scalability
- Decentralised data
- Cloud native and masterless
- A balance of performance and availability
AIM: The data centre market in India is growing rapidly. Do you have plans to leverage this opportunity?
Deb Dutta: DataStax is a cloud-first company. Thus, our focus is more on the rapid growth of India’s cloud services and adoption, which is extremely promising. An IDC report on India Cloud Predictions for 2022 estimates that up to 40% of organisations will implement dedicated cloud services by 2024. Similar reports highlight the remarkable trajectory that India is on where cloud adoption is concerned, and this gives us great confidence as we lay our foundations in one of the fastest-growing economies in the world.
The growth trajectory is being fuelled by the rapid adoption of digital infrastructure amid the pandemic, rising digital usage, cloud consumption and the rollout of 5G in the region.
AIM: What are the key challenges with respect to harnessing real-time data? How is DataStax helping address these?
Deb Dutta: For most operators, scaling, maintaining and updating legacy datastores is an uphill task. The infrastructures of older systems and the support and upgrades they require are very cost-intensive when compared to investing in new software. Moreover, a company cannot detect fraud in real-time or offer personalised experiences on the web at the moment or power machine learning and AI with legacy databases. This can be resolved by moving to the cloud as it helps eliminate maintenance and database infrastructure upgradation costs. Enterprises pay only for what they consume.
Today customer satisfaction is everything to the success of a business. They constantly demand speed and seamlessness in interacting with an organisation’s web and mobile applications. Therefore, adopting data solutions suited for modern applications, including speed, scalability, performance and real-time insights, is critical for enterprises. NoSQL databases like Cassandra, for example, will help accelerate performance and expedite the query response time of some of the most intense cloud applications.
It goes without saying that security is one of the biggest concerns for enterprises today. Legacy systems no longer offer the security that cloud-native applications can, simply because it’s a lot tougher now to get security updates and support from service providers. Most cloud database providers offer extensive services that can help combat cyberattacks.
Another challenge is mobile compatibility. For enterprises to deliver a seamless mobile app experience to their customers, quick access to huge volumes of data is critical. This is where cloud-ready platforms like Cassandra-embed data in the database, which is then replicated in the cloud. Thus, it does away with the need to request data from servers constantly.
AIM: With the growing popularity of NoSQL databases, do you think it can completely take over SQL?
Deb Dutta: NoSQL technology is disruptive. However, it is purpose-built for modern-day applications. Traditional relational database management systems have a purpose – with their obvious limitations. If an organisation deals with projects that require scale and large amounts of real-time data, such as personalisation, user vector, metadata recording, user profile management, content management, mobile applications and recording and inference from IoT sensors, it requires NoSQL technologies to cope with the complexity and volume of data needs.
With massive data and NoSQL growth in the APAC region, modern applications are most needed as digital native companies come on board and legacy enterprises seek to modernise their application stacks. Implementing NoSQL systems into an organisation is a paradigm shift but worth the effort to drive innovation and growth.