On the database market. Q&A with Ariff Kassam
Q1. How is the database market evolving?
The database is moving to cloud; the analysts agree. Today, most of the new database development is going cloud first. That means new features and capabilities will be rolled out in the vendor’s cloud offering first before being available on-premises. While on-premises still has a role (and will continue to do so for some applications), the pace of innovation for on-premises databases will slow down or at least lag behind cloud deployments.
We’re also seeing demand for multi-cloud options. Now that customers are more accepting of security capabilities in the cloud, they are evaluating corporate level strategies. Multi-cloud requirements are being driven by DR strategies, the desire to avoid cloud lock-in, or by regulatory pressures.
Q2. How do you classify the myriad of choices available?
There are a lot of databases available today. In the last few years, the market has really exploded. For operational databases, I use the following classification scheme:
- Non-Relational (non-relational model, BASE, API access)
- Wide column
- Relational (relational model, ACID, SQL access)
- Traditional architectures (single master node)
- On-premises – proprietary vs. open source
- Cloud specific
- New architectures (multi-node/multi-master)
- Cloud specific
- Cloud agnostic
- Traditional architectures (single master node)
Q3. What are your recommendation on how to best align business requirements with the “right” database or “data platform”?
There really isn’t a one-size-fits-all model. All organizations are using more than one database. With all this choice, determining the “right” database or platform has become a multi-faceted optimization decision.
Start with determining the most important technical requirements that the database must support:
- What are the data consistency requirements?
- What are the availability requirements?
- What type of data model is needed? Will it evolve over time or is it relatively fixed?
- What workload does the database need to support? Read/write mix, known vs. ad-hoc queries?
- What are the key performance KPIs? Throughput, latency, concurrency, scalability, etc.
Next consider any business drivers in your organization:
- Is there a mandate to move to the cloud?
- What are the existing skill sets?
- Do you have a need to avoid vendor lock-in?
- Are there regulatory drivers, such as data residency or multi-cloud requirements that must be considered?
By considering the cross section of databases or data platforms that meet both your technical and business requirements, your options for databases should be significantly reduced.
Q4. Who has, in your opinion, implemented a successful (multi) cloud strategy?
There are a number of organizations who are leveraging multiple clouds across their organization. Generally, this is because different development teams are comfortable or need specific services from different cloud providers. The usage of each cloud is isolated and independent of any other cloud. This type of multi-cloud usage is common in many large organizations.
However, there is a FinTech digital bank startup in Hong Kong, WeLab, that is using multi-cloud in an entirely different way. They are deploying a single application across two independent clouds. Essentially, they are using the two clouds for Disaster Recovery to meet regulatory requirements. Together, NuoDB and Rancher Labs will help WeLab deliver a single, logical SQL database across a multi-cluster Kubernetes environment managed by Rancher and deployed on multiple cloud providers, specifically AWS and Azure.
Q5. What is the business value of using microservices and containers in developing applications?
Microservices is an architectural pattern that allows companies to improve their development process to be more productive, agile, and cost effective. It also enables companies to develop a “cloud first” or “DevOps” mentality, allowing development teams to continually roll out new application versions quickly.
One negative side effect of microservice architectures is operational complexity. By definition, converting to a microservices architecture will result in more components needing to be deployed and managed. The benefit of containers and container orchestration systems is to reduce operational complexity for microservices. Containers provide predictable promotion into production as well as automation for easier operations.
Q6. What do you think of the increased Kubernetes adoption? How do you see this adoption developing in the coming year?
Adoption of Kubernetes has really accelerated over the last year and will continue to accelerate in the coming year. Enterprises have seen the benefit of using Kubernetes for stateless applications and are now driving the adoption of stateful applications in Kubernetes. At KubeCon in San Diego, we saw growth in capabilities for stateful K8s applications, increased standardization, and best practices evolving to meet the needs of the Kubernetes ecosystem and community as they grow and mature.
But running stateful applications in Kubernetes is still relatively new. Over the next year we’ll see improved storage capabilities and development of best practices for stateful applications. Another trend that will develop over the coming year is multi-cluster support. As organizations consider running stateful applications in Kubernetes, they want to replicate the state to another cluster for DR protection. Multi-cluster support will enable communication between separate clusters to allow state to be replicated across clusters.
Q7 Anything else you wish to add?
In an always on world, availability is a key requirement for application success. And yet high-profile outages from major cloud vendors persist. To combat the risk of data loss and reduce potential customer impact during such instances, organizations will accelerate adoption of multi-cloud strategies in 2020.
To meet current requirements for continuous availability, many industries today rely on on-premises databases with backups and disaster recovery in a different region in case of emergency. As organizations increasingly move to cloud deployment, there will be a corresponding rise in regulations, particularly for technology-centric industries such as FinTech. Those new regulations will require cloud deployment models that protect consumers from outages from cloud providers.
Ariff Kassam, Chief Technology Officer, NuoDB
Ariff is responsible for defining and driving NuoDB’s product strategy. Kassam brings 20 years of database and infrastructure experience to NuoDB to help the company achieve its vision of a distributed database that can manage an organization’s most valuable data while exploiting the emerging benefits of modern infrastructures such as cloud and containers. Prior to NuoDB, Ariff held senior leadership positions at Teradata, xkoto, and Halcyon Monitoring Solutions.
Sponsored by NuoDB