In HPE OneView, when configuring Blade Servers (typically within HPE BladeSystem enclosures like the c7000), you define networks that serve different purposes depending on their type and configuration. During network creation in HPE OneView, you can assign a network purpose, which determines how the network is used and how it behaves within the infrastructure. The four primary purposes are:
- General
- Management
- VM Migration
- Fault Tolerance
- iSCSI
Here’s a breakdown of each network purpose and its role in Blade Server configuration:
1. General
- Purpose: Used for standard data traffic (e.g., application traffic, user data).
- Use Case: This is the default and most common network type. It carries regular LAN traffic between servers and the rest of the network.
- Blade Context: Assigned to server profiles for standard connectivity (e.g., web, database, or file traffic).
- Notes: No special handling by OneView—treated as a standard Ethernet network.
2. Management
- Purpose: Dedicated to infrastructure management traffic.
- Use Case: Carries traffic for HPE OneView, iLO (Integrated Lights-Out), and other out-of-band or in-band management communications.
- Blade Context:
- Often used to connect server iLO ports (via mezzanine or dedicated management NICs) to the management network.
- Enables OneView to monitor, configure, and manage servers and enclosures.
- Best Practice: Isolate management traffic on a separate VLAN/network for security and reliability.
3. Fault Tolerance
- Purpose: Supports network fault tolerance for high availability.
- Use Case: Used in conjunction with network teaming or failover configurations (e.g., with VMware vSphere, Microsoft NIC Teaming, or Linux bonding).
- Blade Context:
- When creating server profiles, networks marked as “Fault Tolerance” are typically used in redundant uplink sets or teamed connections.
- Helps ensure continuous network connectivity if one physical path fails.
- Note: This purpose does not refer to HPE’s legacy “Fault Tolerance” (lockstep) server technology—it’s about network-level redundancy.
4. iSCSI
- Purpose: Dedicated to iSCSI storage traffic.
- Use Case: Carries block-level storage traffic between servers and iSCSI SANs.
- Blade Context:
- Used when booting from iSCSI or accessing iSCSI-based storage.
- In server profiles, this network is associated with boot volumes and storage initiators.
- Often configured with dedicated NICs or converged network adapters (CNAs) to ensure performance and isolation.
- Integration: Works with HPE OneView’s storage integration features to automate iSCSI boot volume provisioning.
VM Migration
- Purpose: Dedicated to live migration of virtual machines between physical hosts.
- Use Case: Carries traffic for hypervisor-specific migration protocols, such as:
- VMware vMotion
- Microsoft Hyper-V Live Migration
- KVM migration (in supported environments)
- Blade Context:
- Assigned to server profiles to provide high-bandwidth, low-latency connectivity between hypervisor hosts.
- Often isolated on a dedicated VLAN or physical network to avoid interference with production or storage traffic.
- Critical in virtualized data centers where workloads must be moved without downtime (e.g., for maintenance, load balancing, or failover).
- Best Practices:
- Use 10 GbE or higher for VM migration networks to ensure fast, seamless migrations.
- Combine with jumbo frames (if supported end-to-end) to improve performance.
- Do not share this network with general-purpose traffic to avoid congestion.
💡 Note: While HPE OneView itself doesn’t perform VM migrations, it provisions the underlying physical and logical network infrastructure that enables hypervisors to carry out migrations reliably. By tagging a network as “VM Migration,” administrators signal its intended use—helping with documentation, policy enforcement, and integration with orchestration tools.
Updated List of Network Purposes in HPE OneView (for Blade Server Configuration)
| Purpose | Traffic Type | Typical Use in Blade Servers |
|---|---|---|
| General | Standard data | Application/user traffic (e.g., web, database, file services) |
| Management | Infrastructure management | HPE OneView, iLO, server/enclosure monitoring and control |
| Fault Tolerance | Redundant network paths | NIC teaming, failover configurations for high availability |
| iSCSI | Storage (block-level) | iSCSI boot volumes, connectivity to iSCSI SANs |
| VM Migration | Virtual machine live migration | vMotion (VMware), Live Migration (Hyper-V), or similar hypervisor mobility features |
In HPE OneView, when creating or configuring an Ethernet network, you’ll encounter two important settings under the Network Creation tab:
- Private Network
- Smart Link
These settings control network isolation and link-state behavior within the Virtual Connect (VC) domain (e.g., in HPE BladeSystem c7000 enclosures or Synergy frames). Here’s what each does and why they matter in Blade Server environments:
1. Private Network
🔒 Purpose: Isolate server-to-server traffic within the same network
- When enabled, servers connected to this network cannot communicate directly with each other—even if they’re on the same VLAN and subnet.
- Traffic is forced to egress the enclosure via uplinks to an external switch/router and then return (if allowed by external routing/firewall rules).
- This mimics a “port-isolated” or “private VLAN (PVLAN) – isolated” behavior at the Virtual Connect level.
✅ Use Cases:
- Security: Prevent lateral movement between VMs or servers (e.g., in multi-tenant or DMZ environments).
- Compliance: Meet regulatory requirements that mandate workload isolation.
- Testing: Isolate test/dev servers from each other while sharing the same VLAN.
⚠️ Important Notes:
- Only affects downlink-to-downlink communication within the same VC domain.
- Does not block traffic between servers in different enclosures or different VC domains.
- Does not affect communication with external networks (via uplinks).
📌 Example: Two blade servers in the same c7000 enclosure are assigned to a “Private” network called “App-Tier”. Even with IP addresses in the same subnet, they cannot ping each other unless traffic routes out and back via an external gateway.
2. Smart Link
🔄 Purpose: Control network connectivity based on uplink status
- When enabled, if all uplinks for this network (in the associated Uplink Set) go down, Virtual Connect automatically disables the corresponding server downlink ports connected to this network.
- This prevents servers from sending traffic into a “black hole” when upstream connectivity is lost.
✅ Use Cases:
- Failover Clarity: Ensures applications or OSes detect a true network failure (instead of timing out on unreachable gateways).
- Cluster Stability: Helps clustering software (e.g., Windows Failover Clustering, VMware HA) make accurate decisions during network outages.
- Avoid Split-Brain: In storage or database clusters, prevents nodes from assuming they’re still connected when they’re not.
⚠️ Important Notes:
- Smart Link only works if the network is part of a redundant uplink set (or at least has defined uplinks).
- It affects only the specific network—other networks on the same server NIC remain active if their uplinks are up.
- Disabled by default for most networks (except sometimes for Management networks, depending on design).
📌 Example: A blade server uses two networks:
- Production (Smart Link = ON)
- Management (Smart Link = OFF)
If the Production network’s uplinks fail, the server’s Production NIC is logically disabled, but iLO and Management access remain available.
How They Work Together
| Setting | Effect on Server-to-server | Effect on uplink failure |
|---|---|---|
| Private Network = ON | ❌ Blocked (within VC domain) | No direct impact |
| Smart Link = ON | No impact | ❌ Downlinks disabled if uplinks fail |
You can enable both on the same network if needed (e.g., a secure, isolated application network that should also go down if upstream connectivity is lost).
Best Practices in Blade Environments
- Private Network: Use for multi-tenant apps, PCI workloads, or any scenario requiring strict east-west isolation.
- Smart Link: Enable on production data networks and iSCSI networks to ensure clean failover behavior. Often disabled on Management networks to preserve out-of-band access during upstream failures.
💡 Remember: These features are Virtual Connect–specific and only apply to HPE BladeSystem (c7000) or Synergy with VC modules. They do not apply to standard switch-based deployments.
By leveraging Private Network and Smart Link, you gain finer control over security, availability, and failure behavior in your HPE OneView–managed blade infrastructure.