DCIM Automation or Data Center Inventory

DCIM (Data Center Infrastructure Management) is software for visualizing and managing the physical infrastructure of a data center (data center). A lot has been written on the subject on the Internet; various suppliers offer solutions of very wide ranges in price and functionality. But the basis for such solutions is the same: all of them must, at a minimum, store the exact location of the equipment in the racks of the data center and based on location data (and other data) provide their services.

The question of how to enter information about the placement of equipment in racks in DCIM, and I want to highlight in this article. I will not intentionally name the solution providers, but focus only on the technical side of the matter.

Let's start with the easiest way: a data center employee with a notebook goes around all the racks and draws the actual position of all servers, switches and routers, and then transfers the data from the notebook to DCIM. The most advanced version of this method is to walk between the racks from the iPhone and use the special DCIM client (and there really are ones) to drag equipment along the racks with your finger.

The next automation step is to equip the equipment with optically recognizable tags with a barcode, DataMatrix or QR code. It is clear that reading tags using a scanner or smartphone is more convenient and faster than writing in a notebook.

A similar but faster (and more expensive) method uses RFID tags.



The principle of operation is obvious.

All the methods described above do not allow us to understand at what height in the rack (in which unit) the device is installed. But for the correct filling of the DCIM database, this is still necessary. Therefore, we came up with a couple of improvements in the use of RFID tags and a manual RFID reader. One American company proposed installing RFID tags on each unit rack. And when you read the RFID tag of the equipment, at the same time determine the height at which this equipment is installed. The Russian invention offers similar functionality.



In this option, additional RFID tags are installed at the top and bottom of the rack, while the units of the rack must be completely filled (for the data center, the use of blank unit caps is still necessary for proper cooling). The top RFID tag means the beginning of rack scanning, the bottom end, and information on the order of reading tags in conjunction with the height of each piece of equipment (it is obvious that the identifier allows you to get all the data about the equipment) makes it possible to clearly determine the map of the placement of equipment in the rack.
A greater degree of automation, but without accurate localization of equipment inside the rack, can be achieved using a cart with an RFID reader and antennas.



I drove the cart between the racks, and determined what is in each of the racks.

Other methods of inventory of the data center (filling in the DCIM database) practically do not require the participation of personnel. Tracking the movement of equipment between racks is automatic. One of the easiest options is to place your own transmitter in each rack, and equip each equipment with your own transceiver of electromagnetic waves with a unique identifier. Moreover, the transceiver can be an infrared module, passive or active RFID tag, etc. ... It is clear that by changing the radiation power, it is possible to ensure that only those tag identifiers are read that are located near the antenna of a particular rack.



A natural development of the above method is a multi-element composite antenna that emits only within one unit of the rack, which allows you to know the location of each "piece of iron" with accuracy to the unit.



That is, the antenna reads only the identifier of the transceiver of the equipment that is installed in this unit rack.

A similar principle can be applied if you use an electrical contact through the connector. On each equipment install a key-tablet (touch memory, as on intercoms in the stairwell), and on each rack unit install a “tablet” reader connector. When the equipment is installed in a rack, the “tablet” with a unique identifier contacts with the connector, and again, you can get the equipment placement accurate to the unit.



The same result can be obtained in another way using a Russian product. On the rail of each rack unit, you can place sensors that respond to the presence of installed equipment in this unit (busy / unoccupied). This way you can understand where some equipment is installed in the rack. And if the equipment has an ordinary passive RFID tag, then using the RFID antenna you can understand what specific equipment was installed in the units where it worked.



In conclusion, the description of the system that my Russian American friend observed was: “there was a rack with two dozen computers in the network . To determine which particular computer is at which address, the administrator sent a command to open the CD-ROM to this address, and the drive was put forward on the desired computer. Bingo!"

Also popular now: