033383672922-smmma-16654902086099.png
  • Home
  • Our partners
  • Store
  • Services
    • IT Solutions
      • IT Infrastructure Solutions
      • Software Licensing & Cloud Platforms
      • Cybersecurity Solutions
      • Tender-Based IT Solutions
      • Why Our Solutions?
    • App-Web Development
      • Web Development
      • App Development
    • Digital Marketing
      • SMM
      • SEO
      • Content marketing
      • Search Engine marketing
      • Google Advertisements
    • Pricing
      • SMM pricing
  • About
  • Blog
  • Home
  • Our partners
  • Store
  • Services
    • IT Solutions
      • IT Infrastructure Solutions
      • Software Licensing & Cloud Platforms
      • Cybersecurity Solutions
      • Tender-Based IT Solutions
      • Why Our Solutions?
    • App-Web Development
      • Web Development
      • App Development
    • Digital Marketing
      • SMM
      • SEO
      • Content marketing
      • Search Engine marketing
      • Google Advertisements
    • Pricing
      • SMM pricing
  • About
  • Blog
3718-smmma.png
GET IN TOUCH
The Server Is Off. So Why Is Its Management Controller Still Online?

The Server Is Off. So Why Is Its Management Controller Still Online?

Inside an enterprise server is a second computer, with its own firmware, network access, credentials, and considerable authority over the hardware. Research published in 2026 has brought a familiar but often neglected security problem back into focus: who protects the systems used to manage the servers themselves?

 

A Server That Still Answers After Shutdown

 

There is a useful paradox at the heart of enterprise infrastructure: a server can stop running Windows, Linux, or its hypervisor and still respond to remote management commands.

Shutting down the operating system does not necessarily turn off the independent management controller.

As long as the necessary standby power remains available, an authorised administrator may still be able to view hardware status, open a remote console, change certain settings, or power the machine back on.

That is not a malfunction. It is one of the reasons enterprise servers are manageable at scale.

But it raises a question that rarely makes it into a routine operating-system security review:

Who can reach the computer that controls the computer?

The technology is called a Baseboard Management Controller (BMC).

In 2026, security researchers at runZero made it the subject of a project named Lights Out, examining vulnerabilities across multiple BMC implementations and out-of-band management systems.

Early findings were publicised in August, with additional disclosures in September and October.

The point is not that every enterprise server is vulnerable, or that one manufacturer is uniquely at fault.

It is that a server can be carefully protected at the operating-system level while its management environment remains poorly inventoried, inadequately segmented, or months behind on firmware updates.

 

The Second Computer Inside the Server

 

A BMC is a dedicated management processor running separate firmware.

It operates independently of the host operating system, and on many platforms it can remain reachable while the host itself is shut down. Disconnecting all electrical power is, of course, a different situation.

Administrators encounter this technology under familiar product names:

 

Manufacturer Management Technology
Dell Technologies iDRAC
Hewlett Packard Enterprise iLO — Integrated Lights-Out
Lenovo XClarity Controller (XCC)
Other server platforms Various BMCs, including OpenBMC- and AMI MegaRAC-based implementations

 

Depending on the platform, licence, and account privileges, these controllers can offer power operations, hardware sensor readings, event logs, remote console access, virtual media, firmware maintenance, and other management capabilities.

These functions solve real operational problems.

If a hypervisor becomes unresponsive, a remote console may save a trip to the data centre. If the host operating system fails to boot, out-of-band access can be essential for diagnosis.

The security consequence is equally real: the BMC has its own attack surface.

Endpoint detection software running inside Windows does not automatically monitor the controller's firmware.

Patching Linux does not patch iDRAC.

Disabling the host's network connection does not necessarily disable the dedicated management network.

In effect, an enterprise server has at least two environments requiring protection: the workload environment and the hardware management environment.

 

What the 2026 Research Actually Found

 

The Lights Out research should not be reduced to the claim that Dell, Lenovo, and HPE servers can all be compromised in the same way.

The disclosed issues differ considerably in prerequisites, severity, affected firmware, and remediation status.

Three examples help explain why the details matter.

 

Dell: Disabled in the Settings, Still Reachable Over IPv6

 

On October 26, runZero described CVE-2026-65196, a finding involving Dell iDRAC10.

In the implementation examined by researchers, disabling IPMI over LAN did not fully remove network reachability: initial protocol negotiation remained accessible over IPv6.

That does not mean an attacker automatically bypasses authentication.

Valid credentials or an additional vulnerability would still be needed to perform authenticated operations.

Researchers rated this particular issue Low, CVSS 3.7.

They also stated that their twelve Dell iDRAC findings were of low or medium severity, not high or critical.

The assessment was based on firmware analysis and virtualised testing rather than independent verification on physical hardware.

Nevertheless, the case delivers a useful lesson:

A setting labelled “Disabled” is not always proof that the corresponding service is unreachable from every network path.

IPv4 and IPv6 exposure both need to be checked.

 

Lenovo: A Rediscovered Flaw That Had Already Been Fixed

 

In September, runZero investigated an issue in an older build of Lenovo XClarity Controller 2.

Lenovo identified it as a rediscovery of CVE-2024-38508, a vulnerability fixed and disclosed in 2024.

The researchers had been examining XCC2 firmware version 1.10.

Lenovo reported that it could not reproduce the issue on its newer 8.80 firmware, released in 2026.

This is not evidence that current XClarity Controller installations share the same flaw.

It demonstrates why a firmware version matters more than a product name.

An organisation may upgrade its operating systems and hypervisors regularly while leaving the server's management controller on a release that has not been reviewed for years.

 

HPE: Management Authority Magnifies the Consequences

 

On October 13, runZero disclosed several issues involving HPE iLO.

Among them, CVE-2026-65206 affected certain older iLO 2, 3, and 4 implementations.

Under the conditions described by researchers, access to an IPMI service could lead to an administrator session without the corresponding password.

Other disclosed issues concerned SSH and JSON request handling, with some capable of disrupting the availability of management services.

These are distinct vulnerabilities affecting different generations or configurations; they should not be generalised to every iLO product.

HPE published security bulletin HPESBHF05170 with relevant guidance.

Administrators need to determine which generation and firmware build they actually run before choosing a remediation path.

Together, these cases show three very different risks: a configuration setting that does not fully control exposure, an old vulnerability that persists only on unmaintained versions, and security faults in a highly privileged management environment.

 

Why an Attacker Might Value BMC Access More Than an Ordinary Login

 

Most enterprise security programmes rightly devote significant resources to endpoint protection, operating-system patches, identity controls, backups, and network segmentation.

But a compromised management controller can sit outside some of those conventional controls.

Depending on the specific platform and the privileges obtained, unauthorised BMC access could enable someone to:

  • Disrupt service by controlling server power.

  • Use a remote console or virtual media feature.

  • Alter management settings or create accounts.

  • Interfere with firmware maintenance.

  • Exploit the management environment as a foothold for further activity.

These are potential consequences of sufficiently privileged access, not abilities granted by every BMC vulnerability.

Even the narrower prospect of remotely disrupting a critical host is serious for a bank, telecom operator, manufacturing business, or organisation running a virtualisation cluster.

The risk can become greater when many servers, storage systems, and management appliances share an overly permissive administrative network.

 

“We Have a Separate Management VLAN.” Is That Enough?

 

A management VLAN is a useful design component.

It is not, by itself, a complete security boundary.

If routing and access rules allow ordinary workstations to connect to the BMC subnet, a compromised employee laptop could become a path towards hardware management interfaces.

If administrators connect to iDRAC, iLO, or XCC from everyday endpoints, one endpoint compromise may carry consequences for multiple servers.

If management services are exposed directly to the public internet, the concern is even more immediate.

Three levels of separation are worth distinguishing:

1. Logical segmentation.

Management traffic is placed on separate VLANs or subnets, but the effectiveness of isolation still depends on routing and access-control rules.

2. Controlled administrative access.

Management interfaces can be reached only from approved systems, such as secured jump hosts, dedicated administration workstations, or a tightly controlled VPN.

3. Physically separate out-of-band infrastructure.

In particularly sensitive environments, independent switching and cabling can provide additional separation from production networks.

There is no universal architecture for every company.

But the principle is consistent: access to hardware management should not be granted merely because somebody has access to the corporate network.

Dell and Lenovo both recommend reducing unnecessary management exposure.

The 2026 iDRAC10 IPv6 finding also underlines why firewall rules and verification procedures must account for both IP protocol families.

 

Seven Checks Worth Assigning to the Infrastructure Team

 

This does not need to begin as a costly replacement project.

A focused assessment can reveal where the most important gaps are.

1. Inventory every management controller.

Record each server's manufacturer, model, controller type, firmware version, IPv4 and IPv6 management addresses, responsible owner, and business criticality.

Include legacy hardware and servers managed by contractors.

2. Verify network reachability.

Using authorised and appropriately controlled checks, establish whether management interfaces can be reached from the internet, ordinary user segments, or third-party networks.

Review both IPv4 and IPv6.

Avoid aggressive scanning of sensitive hardware without a safe testing plan.

3. Review firmware against vendor advisories.

Treat BMC firmware as a separately maintained product.

Monitor security bulletins and prioritise updates according to actual exposure, vulnerability severity, and the criticality of the affected service.

Test and schedule firmware changes with appropriate availability controls.

4. Audit accounts and privileges.

Eliminate unnecessary shared or stale administrator accounts through approved procedures.

Review contractor access, use unique credentials, and apply least privilege.

Where supported, enable multifactor authentication or enforce it at a secure administrative access gateway.

5. Disable services that are not required.

IPMI over LAN, Redfish, SSH, SNMP, and other management services should have documented operational reasons to remain enabled.

Dell and Lenovo recommend disabling unnecessary IPMI over LAN access.

However, avoid indiscriminate changes: Lenovo notes that disabling IPMI over KCS on some VMware ESXi configurations can interfere with hardware sensor reporting.

6. Collect and review management events.

Track administrative logins, privilege changes, firmware actions, virtual media activity, and power operations where supported.

Forward relevant events to central monitoring, maintain time synchronisation, and make sure that critical management actions can be investigated even when the host operating system is unavailable.

7. Exercise recovery and access procedures.

On a designated test server or during an approved maintenance window, confirm who can access the BMC, who may restart a host, how former supplier access is revoked, and how administrators recover after a management configuration mistake.

Test a fallback access method before tightening controls on critical systems.

 

A Practical Way to Prioritise the Work

 

Not every controller requires the same immediate action.

Observed Condition Recommended Priority
BMC accessible from the internet or an uncontrolled segment Restrict access immediately and review relevant events
Firmware affected by a confirmed critical vulnerability Urgently assess patching and temporary exposure restrictions
Shared administrator credentials or unowned accounts Assign ownership, rotate credentials, and reduce privileges
Properly isolated, supported, updated, and monitored controller Maintain regular reviews and controlled change management

This is an illustrative risk-prioritisation framework, not a regulatory classification or a universal service-level agreement.

 

Why This Belongs in Procurement Specifications, Not Only Security Reports

 

Enterprise server tenders typically focus on processors, memory, storage, network ports, warranty periods, and delivery times.

Those specifications matter, but for equipment expected to remain in service for five years or longer, the management controller's lifecycle deserves equal attention.

Will the manufacturer provide firmware security updates for the intended service life?

Which remote-management security features are included with the licence?

Can administrative access be integrated with the organisation's identity and logging systems?

How will management access work from a disaster recovery site?

Responsibility also needs to be established during implementation.

A system integrator may install a server and provide its iDRAC address to the customer.

But who confirms the initial credentials have been secured?

Who approves the management-network access rules?

Who monitors firmware advisories after the project has been handed over?

If these responsibilities are left undefined, an infrastructure project can move from successful delivery into years of avoidable operational exposure.

For that reason, BMC configuration and documentation should be part of the server acceptance process, alongside hardware testing and warranty verification.

A server should not be considered fully handed over for secure operation until ownership and protection of its management controller are defined.

 

The Least Visible Computer in the Data Centre

 

The BMC is valuable precisely because it remains useful when ordinary administration methods no longer work.

The answer is not to abandon remote management.

It is to treat it as a critical part of the infrastructure, with its own inventory, network controls, patching responsibilities, and audit trail.

The 2026 research does not establish that every iDRAC, iLO, or XClarity Controller is vulnerable.

It does, however, offer a compelling reason to broaden the way enterprises think about server security.

A machine can have a modern processor, an up-to-date hypervisor, endpoint detection, and reliable backups.

Yet its separate management computer may still require attention.

And if that computer has never been included in an asset inventory or security review, the right time to correct that omission is before an incident.

Knowing who can log in to Windows Server is not the same as knowing who can control the server itself.

18.11.2026

  • Home
  • Our partners
  • Services
    • IT Solutions
      • IT Infrastructure Solutions
      • Software Licensing & Cloud Platforms
      • Cybersecurity Solutions
      • Tender-Based IT Solutions
      • Why Our Solutions?
    • App-Web Development
      • Web Development
      • App Development
    • Digital Marketing
      • SMM
      • SEO
      • Content marketing
      • Search Engine marketing
      • Google Advertisements
    • Pricing
      • SMM pricing
  • About
  • Blog
  • Get in touch
  • IT Infrastructure Solutions
  • Software Licensing & Cloud Platforms
  • Cybersecurity Solutions
  • Tender-Based IT Solutions
  • App Development
  • Web Development
  • Privacy Policy
  • Terms of Use
  • Cookie Policy
  • Laptops
  • Desktops
  • All-in-One PCs
  • Monitors
  • Projectors / Printers
  • Smartphones & Tablets
  • TV
  • Video Surveillance Systems
  • Accessories
Subscribe to our newsletter

SMMHub © 2022. All rights reserved.