IPMI remote management: power on and reinstall without a site visit.

What is IPMI out-of-band management

IPMI (Intelligent Platform Management Interface) is the standard for out-of-band server management. The motherboard carries a BMC chip with its own power and network connection, so it keeps running even when the host is powered off or the operating system has crashed. Through it you can control power, read hardware sensors and see what is on the screen.

Every vendor names its BMC differently: Dell calls it iDRAC, HPE calls it iLO, Huawei calls it iBMC and Lenovo calls it XClarity. Toplink DCIM hides those differences. Once a server’s BMC address and credentials are recorded, engineers no longer need to remember each machine’s management IP or install a vendor’s Java client.

What you can do in DCIM

• Power control: power on, shut down and reboot, with forced power-off on supported models;
• Boot device: boot from PXE, a network ISO or the local disk next time, ready for a reinstall;
• Remote console: VNC runs over the BMC channel, so you see POST, BIOS and the whole installation; SSH and RDP run through the operating system;
• Virtual media: mount installation images remotely instead of inserting discs or USB sticks on site;
• Virtual keyboard: send special keys such as Ctrl+Alt+Del and paste clipboard text;
• BMC reset: reset an unresponsive BMC remotely.

Supported servers

Servers are managed through the standard IPMI protocol, whatever the brand. The remote console is tuned for the major BMCs and can detect the model automatically:

Vendor	BMC
Dell	iDRAC 6 / 7 / 8 / 9
HPE	iLO 2 / 4 / 5
Huawei	iBMC, iBMC 6, iMana
Supermicro	Generic, OEM, HTML5 iKVM
Lenovo	ThinkSystem XClarity
Inspur, H3C, Cisco	Inspur OEM, H3C HDM, Cisco IMC
White-box and smaller vendors	Generic AMI MegaRAC

Security and audit

• Console sessions can be shared by link, with read-only or full-control permission and an optional usage limit;
• Every session is recorded and can be replayed at any speed or scrubbed; SSH and Telnet sessions also keep a command-level log;
• Administrators can see who is connected to a device and disconnect extra sessions;
• Customers must enter a one-time password (OTP) before risky actions such as reinstalling or entering rescue mode in the customer portal.

Typical scenarios

• “My server is unreachable”: open the console first to see whether it is stuck at boot, then decide whether to reboot or enter rescue mode;
• Bulk provisioning: set a batch of servers to boot from PXE and let the task queue reinstall them, with progress visible in the admin panel;
• Night-time incidents: on-call engineers can fix issues from home instead of driving to the data center.