Service Fabric (continued)     

Part 2 compared Paxos and Raft. Part 3 discussed SF-Ring, Part 4 discussed its architecture and Part 5 described compute planning and scaling.  This article describes Service Fabric security best practices.

Azure Service Fabric makes it easy to package deploy and manage scalable and reliable microservices. It helps with developing and managing cloud applications. These applications and services can be stateless or stateful. They are run with high efficiency and load balancing. It supports real-time data analysis, in-memory computation, parallel transactions, and event processing in the applications.

The security best practices are described at various levels. At the level of an instance of Service Fabric, the Azure Resource Manager templates and the Service Fabric PowerShell modules create secure clusters. X.509 certificates must be used to secure the instance. Security policies must be configured and the Reliable Actors security configuration must be implemented. The TLS must be configured so that all communications are encrypted. Users must be assigned to roles and Role based Access Control must be used to secure all control plane access.

At the level of a cluster, certificates continue to secure the cluster and client access – both read-only and admin access are secured by Azure Active Directory. Automated deployments use scripts to generate, deploy and roll over the secrets. The secrets are stored in the Azure Key Vault and the Azure AD is used for all other client access. Authentication is required from all users. The cluster must be configured to create perimeter networks by using Azure Network Security Groups. Cluster virtual machines must be accessed via jump servers with Remote Desktop Connection.

Within the cluster, there are three scenarios for implementing cluster security by various technologies.

Node-to-node security: This scenario secures communication between the VMs and the computers in the cluster. Only computers that are authorized to join the cluster can host applications and services in the cluster.

Client-to-node security: This scenario secures communication between a Service Fabric client and the individual nodes in the cluster.

Service Fabric role-based access control: This scenario uses separate identities for each administrator and user client role that accesses the cluster. The role identities are specified when the cluster is created.

A detailed checklist for security and compliance is also included for reference: https://1drv.ms/b/s!Ashlm-Nw-wnWzR4MPnriBWYTlMY6