Browse Prior Art Database

Dynamic HA Redundancy Using TPM Cluster Domain Model and Tivoli System Automation Disclosure Number: IPCOM000124054D
Original Publication Date: 2005-Apr-07
Included in the Prior Art Database: 2005-Apr-07
Document File: 3 page(s) / 69K

Publishing Venue



This disclosure proposes a means that allows IT servers to be dynamically allocated and de-located for the usage as HA (High Availability) backup server. Decisions, when and how servers are allocated or de-allocated are automated. Furthermore we allow that all the process steps are done programmatically as a workflow.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Dynamic HA Redundancy Using TPM Cluster Domain Model and Tivoli System Automation

The user of a system such as an Automation System expresses a certain Hight Availability goal and the right configuration settings enabling this Automation System to take over making the appropriate decisions and taking actions towards these goals.

This Automation System allocates and de-allocates IT servers, considering the following conditions:

1. Is there enough HA redundancy, supporting the defined HA goal?
2. Is the IT application important enough to request additional IT servers? This resumes IT resources can be shared between applications and demands.
3. Are there 'free' IT servers available which can be allocated for this application? For example DB2 UDB requires a Linux x Blades with a FastT 700 storage attachment.

Or, in other words, IT Server (de)allocations are done by finding an optimal solution between constraining conditions:
1. How many IT Servers are available for the different Applications in an enterprise due to cost limitations? 2. The HA objectives, enterprises want to support 3. The importance of some resource. 4. The current situation of whether things are available or can be started reflecting the current operation status of all managed IT Resources

The configuration and installation of the required hardware and software is automated (provisioned) as well and pre-tested. This provisioning automation is done using a workflow engine, orchestrating the different installation and configuration tasks. Changes made in the installation and the configurations are picked from a master copy, after testing. All this is automated following a predefined model, to greatly reduce the probability of configuration and installation errors.

The Automation System is basically an interconnection of following concepts.
1. The concept of automating the (de)allocation of IT Resources like Servers, Network and Storage for a certain IT application, as described and implemented by the Tivoli Provisioning Manager
2. The concepts of a clustered HA Manager as described and implemented by the Tivoli System Automation
3. The concept of an arbiter, that can find an optimal solution between different requests coming from different HA clusters or even different management domains. This is described and implemented by the Tivoli Intelligent Orchestrator

Only the interconnection of these three basic concepts provides the ability to dynamically manage HA goals. The overall Automation System is the interaction between the three components as shown in the following diagram.


Page 2 of 3

1. The HA Manager (Tivoli System Automation) detects that a certain HA goal cannot be achieved without allocating additio...