Storage Management:
Managing the storage supply chain leverages many of the same concepts and practices as traditional supply chain management (Table 1). The key to these concepts is to understand the process necessary to complete the task, whether manufacturing a widget or supplying high available storage to a critical application.
Managing the IT infrastructure and the availability of the IT components like other components of the business is the foundation for managing storage as a supply chain. The goal is to achieve "just-in-time" delivery without sacrificing service levels or increasing costs.
Another perspective on management of supply chains is to consider an initiative that brought technology "closer" to the end user. The Demand Activated Manufacturing Architecture (DAMA) Project began in 1994 with a holistic view of the soft-goods supply chain (fiber, textile, sewn products, retail, and consumer). The DAMA researchers proposed that if Cooperative Business Management could be agreed upon among the companies in the soft-goods supply chain, a higher degree of synchronization would occur. With more synchronization, the time from raw materials to consumer in a supply chain could be reduced by 50%. This would place the decision making of what to manufacture closer to the consumer's decision of what to buy and all participants would benefit. To enable this collaborative supply chain, the researchers developed the DAMA Information Architecture (Chart 1).
In simplistic terms, the DAMA Information Architecture is targeted at driving business decisions further into the infrastructure and driving technology decisions closer to the business. This transformation, tying businesses and IT decisions together, is what the storage management industry is moving toward for fast and measurable ROI of the networked storage environment.
Storage Supply Chain Management
The storage management and supply chain consists of every component within the storage infrastructure that is utilized to keep applications in compliance with performance, availability and cost service level objectives. Storage Operations Management software will automatically, based on user-defined policies, "processize" each action and optimize each component in the storage supply chain to support the application (Chart 2).
This includes not only the application and the server(s) it resides on but the host bus adaptors, the switches and the ports and paths associated with these switches, the storage arrays and the spindles on the arrays. It also includes the logical software components as part of the supply chain: file systems, databases, volume managers, etc. To optimize a storage management supply chain based on business level applications, a policy engine is required to be able to maintain and continually balance the number of and potentially conflicting SLOs of the entire networked storage environment.
Furthermore, the decisions based on the policy and rules must be coordinated across the distributed resources, people and technology involved in the delivery. This requires integration of the policies and derived actions into a workflow that controls the distribution and control of those tasks to the distributed resources.
It is important to enable the levels or tiers of storage that were defined in the production step to determine what suppliers or resources will be able to achieve these. Elements of the supply chain need to be discovered and characterized as to their role in delivering the desired storage service. It is critical that resources in the supply chain are capable of delivering the expected service levels and cost objectives of the service. Storage management software should be able to analyze these costs and service levels as part of managing the supply chain.
Key to the storage management processes is determining how much inventory is available for new orders. In a storage supply chain, inventory is both disk capacity and bandwidth in the network. This requires understanding the following:
1. What is allocated to existing customers and applications?
2. How much is already configured and available?
3. How much is available to reclaim from existing customers?
4. How much is in a raw or unconfigured state that can fulfill forecasted?
5. What is the backlog of requests and forecasted demand?
Location
The location for this storage or storage environment must be considered next. Where are critical applications being hosted and accessed from? Should we have multiple networked storage environments? Where should the backup or redundant storage be located? What are the limitations on latency and performance based on distance?
