Data Management
October 10, 2022 18:04:46 UTC
Last Updated
This Standard Operating Procedure (“SOP”) document establishes the Data Management policies and procedures pertaining to the services provided by Stellar Technologies Inc (“Stellar”).
RPO & RTO
The following illustrates the difference and impact of RPO (Recovery Point Objective) and RTO (Recovery Time Objective).
Production
Production data storage equipment meets the following specification:
- Dual Redundant Controller
- Raid 50 or 60 Configurations consisting of:
- 2x 12 Drive Data Sets
- 2x Global Spares
- Redundant Volume Remote Copy to archive array for 24-hour snapshot redundancy.
Encryption
Customer-defined "production" data is encrypted at rest using AES-256 encryption via either self-encrypting drives or VMware Virtual Machine Encryption.
Isolation
Data is isolated on a per-customer basis by volume(s) and/or Logical Unit Number(s) (LUN). No two customers ever use the same storage volume or LUN.
Performance
Customer-defined "production" data resides on all-flash storage arrays. Storage IOPS policies are implemented to guarantee customer performance.
Backups
Customers may elect for backups of any offering provided by Stellar. Backup services are consumed based on the number of machines protected and the total aggregate storage consumption.
Data backups may be stored in a Local Redundant Zone and/or a Global Redundant Zone:
- Local Redundant Zone: Stored in the same data center as production systems, but on different equipment.
- Global Redundant Zone: Replicated to opposing regional data center.
Data backups are stored on high-performance enterprise storage to ensure the lowest possible customer data restoration times.
Standard RPO & RTO
Stellar utilizes the following RPOs and RTOs by default:
RPO | RTO |
|---|---|
24 Hours | Up to 1 TB per hour |
Customers may define custom RPO & RTO targets if needed.
Encryption
Backups are encrypted at rest via Veeam Encryption Keys. Customers subscribed to backup services may set their own encryption key, or use a randomly generated key provided by Stellar.
Stellar-assigned keys are created on a per-tenant basis, and no key is reused. Automatically rotated keys are available upon request.
Encryption Keys
Stellar is not capable of restoring data without a customer's encryption key. Customers are strongly advised to securely store their backup encryption keys in more than one place.
Isolation
Data backups are isolated on a per-customer basis by volume(s) and/or Logical Unit Number(s) (LUN). No two customers ever use the same storage volume or LUN.
Performance
Backup restoration from source data located within an Orion facility to a target within the same Orion facility has been proven to perform at a rate of at least 1 TB per hour.
Backup restoration from source data located within an Orion facility to a target within a different Orion facility can perform at variable rates, due to differences in available bandwidth or round trip time, depending on the facilities in question. In general, customers can expect a performance rate of at least 400 GB per hour in this scenario.
Backup restoration from source data located within an Orion facility to a target located at the customer premises can also perform at variable rates, depending on the available bandwidth at the customer premises, round trip time, and other congestion factors.
Notifications
Customers must provide an email address for the purpose of delivering backup success/failure notifications. To request backup notifications, please open a support case.
Recovery
The maximum potential loss of data for Stellar-initiated data recovery events as a direct result of hardware failure is 24 hours (the maximum amount of time between volume copy jobs), and the minimum is 1 hour. Customers impacted from any data recovery event are subject to the maximum credit per Service Level Agreements.
Disaster Recovery
Replication
Customer VMs may be replicated asynchronously with a maximum RPO of 1 hour, and will be ready to power on within the customer's Orion cloud environment.
Performance
Disaster Recovery as a Service customers can expect near real-time recovery by establishing a virtual standby instance of critical VMs in their DR environment.
Failover Testing
Customers may conduct up to two failover tests per month, as long as the total compute runtime does not exceed two hours.
Declaring a Disaster
Customers may declare a disaster at any time. Disaster declarations must be provided to Stellar via Phone or Email by a pre-defined authorized contact.
A customer's environment may run in the failed over state for up to 30 calendar days at no additional cost to the customer. If the environment needs to operate in a failed over state for more than 30 calendar days, the customer will be charged a predefined rate per day, per VM.