Overview
Why Use Accelerator?
Vertica Accelerator offers Vertica as a service (SaaS), allowing you to leverage Vertica’s blazing speed without worrying about the intricacies of database setup, administration, or maintenance.
Accelerator’s fast and easy database deployment and management options allow you to do the following with just a few clicks:
- Create and manage database
- Administer (start, stop, resize, drop) subclusters
- Dynamically scale subclusters
- Set up idle shutdown
- Set up backup and restore
Accelerator Architecture
Accelerator allows you to dynamically scale clusters to meet your workload. Using Vertica Eon Mode, Accelerator increases or decreases your clusters to meet your workload while saving money.
- Frequently accessed data is stored in the depot (SSD drives attached to the vCPU) to increase access speed, while less-accessed data remains in the S3 communal storage to save costs.
- Multiple subclusters are assigned as dedicated resources, allowing workloads to be isolated to ensure a bottleneck on one cluster does not affect otyher clusters.
- Compute resources can easily be increased or decreased by changinng the number of nodes and the compute capacity of each node (EC2 instance size).
- Scaling options are:
- Scheduled autoscaling: Schedule your subcluster(s) to start and stop at specific times; ideal for when you have a consistent workload.
- Elastic autoscaling: Configure additional subclusters to start when your real-time workload exceeds a certain threshold and automatically stop subclusters when workload is low. This option enables you to balance cost and performance when the workload is not predictable enough for scheduled autoscaling.
Vertica in Eon Mode
Accelerator runs Vertica in Eon Mode, allowing you to separate storage and compute functionality.
Eon Mode separates the computational process from the communal storage. This enables you to scale up or down according to your workload, by adding or removing nodes, as well as by increasing or decreasing instance size on the fly.
Vertica in Eon Mode is suited to a range of needs and data volumes. Because compute and storage are separate, you can scale them separately.
Since Vertica Accelerator utilizes Vertica in Eon Mode, you can create subclusters within your database to isolate work. For example, you might want to dedicate some nodes to loading data and others to executing queries. Or you might want to create subclusters for dedicated groups of users (who might have different priorities). You can also use subclusters to organize groups of nodes for easily scaling your database cluster up and down.
Anything accessible via SQL is supported in Accelerator. Anything requiring Management Console or OS access is not part of the Accelerator user interface, but the Accelerator support team may be able to help customers via support to load things on the cluster if necessary. ISO is supported, SOC II is in process, and PCI does not apply.
Use Case 1: Volatile/unpredictable workload
Use elastic autoscaling and connect everything to the autoscaler DNS. Elastic autoscaling automates starting or stopping additional subclusters according to your real-time workload.
Use Case 2: Predictable workload
If your analysts are typically active during office hours, use schedule-based autoscaling in conjunction with workload isolation by subcluster, where ETL, analytics, reporting, are each given a separate subcluster.
In either case, it is recommended that you enable idle shutdown to optimize for cost-savings.
To get access to try out Vertica Accelerator, register at https://www.vertica.com/landing-page/accelerator/, and a member of the Accelerator team will be in touch.