You are here

NuoDB Multi-Tenancy Demonstration

Multi-Tenancy

This short video demonstrates NuoDB's unique take on multi-tenancy. The video covers deploying, managing, and scaling databases on shared resources from NuoDB's Administration Console.

Video transcript: 

NuoDB is a scale-out SQL database system that allows you to dynamically add and remove active nodes without ever sacrificing transactional consistency.

In this Demo video we will showcase NuoDB’s unique take on Multi-tenancy. 

NuoDB allows multiple databases to share resources and be managed through a single administrative interface. This pool of shared resources is called a Domain. 

Within a Domain, you may have computing resources available to you across a number of hosts. 

Each database (Begin showing Databases) may use resources across one or more hosts but remain independent of the other tenants, maintaining its own storage and security credentials, despite sharing computing resources. 

All resources available for use by the databases within a Domain are managed by NuoDB’s orchestration capabilities.

In addition, a single administrative interface manages all of the databases.

This architectural set-up enables NuoDB databases to run processes on shared host resources to maximize server utilization in deployments. Dependent on application usage, NuoDB can run thousands of active databases across a few dozen hosts.

Now let’s switch over to the NuoDB automation console to drill down further into NuoDB multi-tenant capabilities.

As you can see from the Dashboard we currently have four active single host database. Each database is self-contained on it’s own host, but all metrics and reporting roll up to the NuoDB Automation Console dashboard. 

As we select one of the hosts we see that the Storefront3 database is running 2 processes, a transaction engine or TE to handle transactions and a storage manager or SM to make commits durable on disk.

Each storefront database is currently serving as the database for an e-commerce application with 100 concurrent users browsing, purchasing items and writing products reviews. Each database is currently processing around 240 transactions per second at a latency of 6 milliseconds per transaction.

Now lets assume we’ve recently acquired a new E-commerce tenant that has 500 concurrent users of their application. As the new storefront5 database comes online we can see that is starts up its own unique TE and SM processes on the host where the first Storefront database is running. 

As you can see the Storefront5 database has unacceptably high latency which is why we are going to scale out its database resources. As we change the slider from 1 to 4 hosts the database now automatically applies NuoDB’s multi-host template. This template (Show Icon and definition) starts a second SM process and a TE process on all available hosts in the cluster. As you can see on the screen the Storefront5 database now has 6 processes running across 4 hosts.

This increase in resources has given the Storefront improved throughput and a 10x decrease in Average Latency per transaction in a matter of seconds. All without any degradation in performance of the other databases in the cluster.

What we’ve seen in this video, in real time, is NuoDB’s ability to easily deploy and manage a multi-tenant database setup through a single management interface. Adding or removing tenants or scaling out a tenant’s database can be accomplished on demand.