On-Premises PCE Design and Architecture

Datasheet

Overview

The On-Premise PCE Design and Architecture add-on gives customers a right-sized design for an on-premise Policy Compute Engine before any infrastructure is built.

An on-premise PCE runs on one or more nodes. The deployment type depends on scale and resilience requirements. A 2x2 multi-node cluster runs two core nodes and two data nodes for redundancy. A 4x2 cluster runs four core nodes and two data nodes for larger deployments. A supercluster joins two or more replicating clusters into one administrative domain. If the wrong type or size is chosen, the PCE can be under-provisioned, lack resilience, or be expensive to rebuild later.

Illumio Professional Services works with the customer to determine the right PCE type and scale for the current deployment and planned growth. PS then specifies the resources each server needs, defines where the nodes should be placed, and gives the customer setup instructions so the environment is prepared correctly. If the customer plans to split the PCE across data centers, PS helps define the latency and connectivity requirements between sites.

This add-on covers preparation and design only. The PCE itself is built in the separate On-Premise PCE Build add-on.

At the end of the engagement, the customer knows what needs to be built, how large it must be, where it should be built, and what infrastructure must be ready before installation starts.

Prerequisites

Package Dependencies

Segmentation Strategy & Roadmap
This add-on is normally followed by the On-Premise PCE Build add-on.

Technical Readiness

Decision to deploy the PCE on-premise
Estimated current and planned workload count
Candidate data center locations identified Information on available server, virtualization, network, and storage infrastructure

Customer Readiness

Project manager or primary point of contact assigned
Infrastructure, network, and data center stakeholders available
High availability, disaster recovery, and growth requirements identified
Existing infrastructure standards documented, including server builds, load balancing, certificates, DNS, and email

Covered Services

Requirements and Capacity Review

Review the current workload count, planned growth, and expected scale of the deployment.
Review high availability, disaster recovery, and resilience requirements.
Review candidate data center locations and the customer's infrastructure standards.
Review integration requirements that affect the design, including load balancing, certificates, DNS, and email.
Apply Illumio capacity planning guidance to the customer's requirements.

Customer Outcome

The customer's scale, growth, and resilience requirements are documented and ready to be turned into a PCE design.

PCE Architecture Design

Recommend the PCE deployment type: single-node cluster, 2x2 multi-node cluster, 4x2 multi-node cluster, or supercluster.
Specify the resources required for each node.
Define node roles and placement.
Where the PCE will be split across data centers, define node placement, the latency requirements between sites, and the connectivity requirements.
Define the load balancing approach for multi-node clusters.
Define requirements for certificates, DNS, email (SMTP), and network and firewall access.
Define backup, high availability, and disaster recovery considerations for the recommended design.
For a single-node cluster, this includes the additional backup and recovery precautions Illumio recommends because an SNC is a single point of failure.

Customer Outcome

The customer receives a right-sized PCE architecture that sets out what will be built, where it will be built, and how it will scale and recover.

Mitigation Planning

Build a mitigation plan for each confirmed cross-boundary finding.
Classify each finding as:
Traffic to allow permanently
Traffic to allow temporarily, pending remediation
A label correction
An infrastructure change needed, such as moving a system out of the wrong environment, data center, or VLAN Document the owner, rationale, and recommended action for each finding.
Record temporary exceptions so they can be tracked and removed later.
Agree the mitigation plan with customer stakeholders.

Customer Outcome

The customer has an agreed plan for every cross-boundary dependency, so separation can go ahead without blocking legitimate production traffic.

Separation Policy Development

Develop environment separation policy for the agreed boundaries.
Build approved permanent and temporary exceptions into the policy.
Validate the policy against the existing core-services and ransomware containment framework. Organize the policy structure and related policy objects. Create the separation policy in draft so the customer can review it before provisioning.

Customer Outcome

The customer receives draft environment separation policy that restricts cross-environment traffic while preserving approved dependencies, ready for customer review and provisioning.

Environment Preparation Guidance and Handoff

Provide PCE design documentation covering the deployment type, node specifications, placement, and supporting requirements.
Provide setup instructions the customer can use to prepare servers, networks, and supporting services before the build.
Provide a readiness checklist to confirm the environment is prepared for installation.
Document design assumptions, dependencies, and decisions still open with the customer.
Review the design and setup instructions with customer stakeholders.

Customer Outcome

The customer has the documentation and instructions it needs to prepare its environment correctly, so the On-Premise PCE Build add-on can start without redesign or rework.