Site News  |   Monitor Keywords  |   Monitor Archive  |   Organizer  |   Account Info  |

Systems and arrangements to adjust resource accessibility based upon usage modes

Pursuant to 35 USC § 120, this continuation application claims priority to and benefit of U.S. patent application Ser. No. 11/070,053, entitled “SYSTEMS AND ARRANGEMENTS TO ADJUST RESOURCE ACCESSIBILITY BASED UPON USAGE MODES”, attorney docket number AUS920050130US1(4107), filed on Mar. 2, 2005, the disclosure of which is incorporated herein in its entirety for all purposes.

The present invention generally relates to the field of managing access to resources oil a networked computer system. More particularly, the present invention relates to systems and arrangements to adjust resource accessibility based upon usage modes such as the accessibility of resources on an InfiniBand™ network to a supervisor instance based upon a usage mode associated with the supervisor instance and the resources of that supervisor instance makes available to clients of that supervisor instance based upon the usage mode of the client.

Many different types of computing systems have attained widespread use around the world. These computing systems include personal computers, servers, mainframes and a wide variety of stand-alone and embedded computing devices. Sprawling client-server systems exist, with applications and information spread across many PC networks, mainframes and minicomputers. In a distributed system connected by networks, a user may access many application programs, databases, network systems, operating systems and mainframe applications. Computers provide individuals and businesses with a host of software applications including word processing, spreadsheet, accounting, e-mail, voice over Internet protocol telecommunications, and facsimile, as well as databases, service objects, file systems, I/O devices e.g., data storage, and the like.

Users of digital processors such as computers continue to demand greater and greater performance from such systems for handling increasingly complex and difficult tasks. As a result, demands for fast communications and accessibility to a variety of resources have increased significantly. As an illustration, in some current systems, if a user requires high reliability on some files and fast access on other files, a system administrator may need to allocate storage on different volumes to provide the different storage characteristics. The user may then have to keep track of where the different types of files are located within the network. For example, to obtain high reliability, a file may be stored within a volume corresponding to a data mirroring storage device. Alternatively, a volume corresponding to a data striping storage device may be used for storing a file requiring higher performance.

To satisfy the growing demands for availability of resources while balancing the availability with the costs of purchasing and maintaining large networks, new network models also allow dynamic allocation, deallocation, and reallocation of resources shared by multiple processes operating oil multiple computer systems. For instance, multiple single or multi-processor computer systems may be connected together to create a cluster. Clusters Of multiple systems provide increased compute capability. One such clustered system is a new industry standard architecture, commonly called InfiniBand™. InfiniBand™ is capable of providing decreased latency, greater bandwidth and increased expandability by providing a system-area network (SAN) that includes a channel-based, switched-fabric technology.

Each device or node coupled with networks such as SANs, whether a processor or I/O device, includes a channel adapter. The messages are transmitted from one device's channel adapter to another device's channel adapter via switches. InfiniBand™, for example, provides an industry standard interconnect for clustering and defines a mechanism for message transfer between end nodes. The Architecture outlines key management requirements for initialization, configuration, and control of communications between the nodes connected to the network.

In an InfiniBand fabric, for example, a solution has been proposed which allows resources within an I/O Unit (IOU) to be restricted to a set of hosts. Hosts are identified by their Globally Unique Identifiers (GUIDs). The resources which may be assigned to a host include access to Service Objects, number of Queue Pairs (QPs) reserved for this host, maximum number, of QPs that the host may consume, and the like.

A problem with this approach is that hosts may be used for multiple purposes, and these varied purposes may require different access rights. For example, if a host is normally used as a development and test system, then an installation may wish to restrict access, disallowing access to service objects which represent production resources, and allowing only a limited set of QPs.

However, that same system may also serve as a failover target. For example, in the event that a production system fails, or additional capacity is required for production workloads, the development and test system may run production workloads. In this case the system will need the access rights and resource allotment associated with the production system.

Solutions that require a system programmer to go out to a management console, and redefine the resources so the host could access the production resources when the failover occurs are impractical and problematic. Given that the environment failovers is normally very hectic, the added steps of doing this not only increases the time required to failover, but also provides an opportunity for human error where either this step is overlooked, or not flawlessly implemented.

Solutions that provide rights to the host node for all the possible resources to avoid the problems with reprogramming resource accessibility invite other problems. More specifically, hosts may be used for various purposes, for example a host may normally be used for software development, however it may also be used for production when the primary host fails, or when additional capacity is required. In such an environment, a customer may not wish to allow the host to have access rights to all the same resources when operating in each mode.

Therefore, there is a need for systems and arrangements to adjust resource accessibility based upon usage modes of the host nodes such as the accessibility of resources on an InfiniBand™ network to a supervisor instance based upon a usage mode associated with the supervisor instance and the resources of that supervisor instance makes available to clients of that supervisor instance based upon the usage mode of the client.

The problems identified above are in large part addressed by systems and arrangements to adjust resource accessibility based upon usage modes. One embodiment comprises a method to define a subset of resources available to a supervisor instance to assign to clients operating under the supervisor instance. The method generally involves determining a set of supervisor keys based upon an identifier associated with the supervisor instance; selecting a supervisor key name based upon a usage mode associated with operation of the supervisor instance; and selecting one or more supervisor keys from the set of supervisor keys based upon the supervisor key name, the one or more supervisor keys to verify authorization for the supervisor instance to obtain a set of client keys, wherein each client key of the set of client keys, when attached to a request by a client of the clients, provides verification of an authorization for the client related to establishment of a connection, the connection being associated with a resource of the subset of resources.

Another embodiment comprises a method for defining access by a client to a subset of resources available to a supervisor instance. The method generally involves determining, by the client, a client key name associated with a usage mode for the client; requesting a client key from the supervisor instance based upon the client key name; receiving the client key from the supervisor instance, the client key being associated with the subset of the resources; and associating the client key with a request to establish a connection with one of the subset of the resources to provide a verification that the client is authorized to access the one of the subset of the resources.

Another embodiment comprises a method for defining a subset of supervisor keys available to a supervisor instance. The method generally involves associating one or more supervisor key names with one or more unique host identifiers; and selecting one or more supervisor keys of the subset of supervisor keys to associate with each of the one or more supervisor key names, wherein the association of the one or more supervisor keys with a supervisor key name of the one or more supervisor key names, defines a subset of the resources that are available to the supervisor instance with the supervisor key name.

Another embodiment comprises a system, which may comprise a configuration to define tables that associate supervisor keys with client keys and with unique host identifiers of nodes to couple with a communications network; a configuration manager coupled with the configuration to define authorized connections between the nodes based upon associations in the tables to prevent unauthorized connections, wherein the configuration manager is adapted to associate one or more supervisor key names with one or more of the unique host identifiers and select one or more supervisor keys of the supervisor keys to associate with a supervisor key name of the one or more supervisor key names, wherein association of the one or more supervisor keys with the supervisor key name defines a subset of resources that are available to a supervisor instance.

A further embodiment comprises machine-accessible medium containing instructions, which when executed by a machine, cause the machine to perform operations for defining access by a client to a subset of resources available to a supervisor instance, comprising determining, by the client, a client key name associated with a usage mode for the client; requesting a client key from the supervisor instance based upon the client key name; receiving the client key from the supervisor instance, the client key being associated with the subset of the resources; and associating the client key with a request to establish a connection with one of the subset of the resources to provide a verification that die client is authorized to access the one of the subset of the resources.