can recover automatically without human intervention in the event of a software failure, without the expectation of continuous uptime. This is built directly into the software, within each individual component or container. Fault tolerance is supported by running the
Resonate
containers with Kubernetes. If a component fails, it automatically restarts and picks up where it left off. Data queues are buffered to allow the recovering processes to continue without losing critical data. Fault tolerance is supported in both single-node and multi-node configurations.
High availability adds the ability to increase overall uptime by distributing the software operation across multiple hardware nodes. If a node has a catastrophic hardware failure, the overall system can continue to operate.
Resonate
must be installed in a multi-node configuration to support high availability.
When operating in high availability mode, the system ensures that the software and the data being processed are replicated across the multiple nodes. High availability coordination and replication require their own processing and networking, so some of the increased processing power from the additional nodes is used for that purpose. All the nodes share the data being processed, and replicate copies of the software and the databases. Nodes mostly run the same code, but some of the databases use a primary and secondary architecture. The software automatically ensures that one of the nodes is primary and the others are secondary.
In the event of a failure of a secondary node, the system continues to operate at reduced net reliability. In the event of a failure of the primary node, the system automatically negotiates which secondary node becomes the new primary, reconfigures it as necessary, and resumes operation. Operations other than database reads and writes, such as RFID reading, network data queueing, and algorithm processing, mostly continue while the database operation recovers. Depending on the specific failure (for example, power off, hard disk removal, or network disconnection), the recovery can take from no time up to a few minutes.
High availability mode requires multiple nodes and further requires an odd number of nodes. The
Resonate RFID Reader Management
Software Installation Guide describes how to set up a three (3) node system. It is possible to install a system using more nodes (for example, 5 or 7 or more) to get higher reliability and availability (but not higher capacity). If this is required, contact your Zebra salesperson, who can help you work up a special project and statement of work with Resonate product management and Professional Services.
Resonate
is currently offered only for on-premises or customer cloud deployment.
Resonate
is only part of a system solution consisting of the customer's computers, network, and other solution software. Due to these factors, Zebra cannot offer a specific service level agreement (SLA) or predefined system uptime percentage. Zebra performs long-term testing and will provide updates on our observed reliability as and when possible.
