Browse Prior Art Database

Automated, User configurable naming algorithm for servers in dense environments

IP.com Disclosure Number: IPCOM000099012D
Original Publication Date: 2005-Mar-09
Included in the Prior Art Database: 2005-Mar-09
Document File: 2 page(s) / 24K

Publishing Venue

IBM

Abstract

As technology progresses, each generation of servers becomes smaller and more powerful. Server installations become increasingly denser. Initially the smallest form factor in the Intel space was a single rack unit (1U). This resulted in a limitation of ostensibly 42 servers per rack. With the introduction of Blade Servers, it becomes possible to have individual servers, which require less than a single rack unit of space. For example, the IBM BladeCenter unit is 7U high and accommodates 14 individual Blade Servers. Thus, it is possible to install 82 servers in a rack. When using system management software, administrators need to be able to correlate the representation in the application with the physical server itself. Administrators have their own rules or naming convention that they use to identify the server in their system management application. As the environment becomes denser and administrators are managing more and more systems, it is increasingly harder for them to identify individual servers. In addition, the ease with which the servers can be moved around, applying their naming convention manually to each representation is a tedious and time-consuming process. Not only does it need to be done when the server is first discovered by the management application, it also needs to be done if one of the values used to generate the name changes. For example, in the BladeCenter Chassis environment some administrators may want to name their blade servers after the bay number of the chassis, the chassis and rack in which the server is located. When the management application discovers the server, the administrator will give it a name, for example Rack 2 Chassis 4 Bay 7. If the blade server is moved to a different chassis and a rack, the user needs to manually change the name to a new value, for example Rack 37 Chassis 1 Bay 5. This article describes a solution to this problem through automation of the naming process. Using a simple grammar based on substitution strings, the users can create rules for how they want names to be generated and assigned when the system management application discovers a new server.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Automated, User configurable naming algorithm for servers in dense environments

   The core idea described in this article is to allow system administrators to define a naming convention, to be used in a system management application, based on a list of pre-defined system descriptors / identifiers applicable to specific components of the dense server environments (i.e., blades, chassis, racks, etc.). Consider the example of the blade servers. In addition to the conventional data like machine type model, FRU (Field Replaceable Unit), fields from the VPD block, etc., one could use the chassis bay location, bay range, name of the chassis it belongs to, IP address, OS type, or system name as the system descriptors / identifiers. The system management software will then allow the system administrator to define a name template made up of one or many combinations of these system descriptors / identifiers along with user-specified strings in any particular order preferred by the administrator. These naming convention templates could be defined before the discovery of any systems and be used to name all the systems during the discovery process or these templates could be tailored after the discovery of systems to personalize a certain set of systems. In either case of naming or changing the system name, the user-specified string from the template remains unchanged and the system descriptors / identifiers get replaced by the values applicable to specified system.

   Currently, most of the system management applications use a fixed naming convention, which gets applied to all the discovered systems, although, post-discovery modification of system name is allowed on an individual system basis. The proposed invention allows the system administrator to define the naming convention and selectively apply it, and it also provides the ability to automate the naming process.

   In the IBM Director system management software, this solution will allow uses to name / rename the managed objects used to represent the various components of the dense environment based on a naming convention chosen for each kind of component. These names are in effect only in the system management software environment and do not affect any hardware settings. The invention works by first defining a list of all possible system descriptors and identifiers as keys that can be used to represent each of the specific components in the dense server environment. Examples of components would be blade, chassis, rack, etc. If at a latter...