Support network robustness, and ensure that the Node Operator is not
completely compromised by a bug in a single client implementation.
Audit Procedures
Check what clients are available for each Protocol, and verify that
where multiple clients exist, the Node Operator has a diversified fleet
of clients deployed across nodes.
Key Evidence
Node deployment inventory listing client software per node
Client diversity policy/strategy document
Configuration management records (such as Ansible, Terraform,
Kubernetes manifests) showing multi-client deployment
Monitoring dashboards showing client distribution across fleet
Success Criteria
Operator deploys multiple client implementations where available for
a protocol
Deployment inventory is accurate (confirmed by observation or system
evidence)
Operator can articulate a client diversity strategy and demonstrate
awareness of alternatives
Failure Criteria
Any of these circumstances is likely to mean the Requirement is not
met.
Single client for a protocol where multiple production-ready
implementations are available, with no documented justification
Deployment inventory is inaccurate or does not exist
Operator is unaware of available alternative client
implementations
Justifiable Exceptions
There are several reasons that justify a Node Operator running a
single-client operation where multiple clients are available,
including
Performance issues for specific clients
Resource constraints
Complexity of managing multiple pieces of software effectively and
securely
Failure to deploy multiple clients where they are available does need
to be justified.
Common Failure Patterns
The Node Operator is aware of alternatives but has never deployed
more than one;
A secondary client is deployed, but not maintained;
There is client diversity at one layer but not another (e.g.
consensus clients but not execution clients)