The Data Management Forum plans to clear the confusion surrounding information life-cycle management.
By Michael Peterson and Edgar St. Pierre
It's clear that information life-cycle management (ILM) holds great promise for end users. However, exactly what ILM is and how the industry will get there have been hotly debated for more than two years. In response, the Storage Networking Industry Association's (SNIA's) Data Management Forum (DMF) has developed a common definition for ILM and an industry road map on how to get there.
ILM is a new management practice for data centers; it is not a specific product, nor is it just about moving data to low-cost storage resources. It is a standards-based approach to automating data-center operations by using business requirements, business processes, and the value of information to set policies and service level objectives (SLOs) for how the supporting storage, compute, and network infrastructure operates.
The key question that flows from this vision of ILM is, "How do we get there?" since these capabilities don't fully exist today. SNIA's DMF plans to unify the industry toward a common goal, develop standards, facilitate interoperability, and conduct market education around ILM. This effort begins with a shift from storage management to data and information management-the lifeblood of today's business processes. Over time, this focus will expand by leveraging and cooperating with other industry groups to relate the management of infrastructure services to data and information. To fully realize the benefits of ILM, this development process requires a road map to guide the way.
According to the SNIA, "Information life-cycle management is comprised of the policies, processes, practices, and tools used to align the business value of information with the most appropriate and cost-effective IT infrastructure from the time information is conceived through its final disposition. Information is aligned with business requirements through management policies and service levels associated with applications, metadata, and data."
The ILM framework illustrates the definition. Goals Management abstracts capabilities from the IT infrastructure in terms of policies and practices, and rationalizes those against business requirements and information value (see figure, above). This allows for automation of the IT infrastructure and associated services.
A critical goal of this ILM vision is a new set of data-centric standards based on SNIA's Storage Management Interface, or SMI. The resulting automation sounds simple, but will require years of work to achieve heterogeneous interoperability.
ILM road map
SNIA's DMF has developed a road map to guide the deployment of ILM along a path that starts with best practices using today's solutions and leads toward ILM-enabled practices using future solutions (see figure, below). The road map begins with the need to establish a strong foundation and culminates with fully automated, multi-site, heterogeneous "enterprise ILM." The road map is divided into five phases, each describing a distinct step toward an ILM-based operating environment, that can be used to describe where in the process a particular data center's implementation stands.
Phase 1: An ILM-based platform requires a strong foundation. The recommendation is to first consolidate your storage and data services around network storage. Reduce the variety and begin standardizing the configurations you are using. In other words, get lined up to move to Phase 2 by preparing the foundational architecture upon which you can build ILM-capable services. Define your data-center architecture and requirements.
Phase 2: The objective of Phase 2 is to standardize your data and storage services so that they are consistent, repeatable, and efficient processes. This begins with a data classification effort. Whether it is conducted for a single line of business or for the entire company, it is necessary to understand the requirements of existing and planned data first. This should include a ranking of applications and their data, relative to one another, based on the value of that information to the business. This value can include intrinsic value based on how much money is lost due to unavailability of the data or the application, and it can include extrinsic values such as risk associated with loss of the data.
The value of the data then leads to the definition of SLOs-acceptable levels of service for performance, availability, operational recovery, disaster recovery, security, and more.
By associating the applications and data with standardized configurations that achieve the desired SLOs, you have a better opportunity to associate appropriate cost-performance configurations to your applications. Associating the value of information with the appropriate infrastructure and services is the guiding principle of ILM. And with consistent configurations, you have a better chance of achieving consistent, predictable, behavior at lower costs. Now your process begins to scale.
Today, many companies conduct classification engagements with professional services consultants who can help you through this potentially complex process. If you begin this yourself, keep it simple. Understand and rank the value of applications and data to your company and match that against an inventory of existing and needed services.
The most crucial step here is to get buy-in from across the company. Seek cooperation from legal, compliance, IT, and other business units.
This process will be simplified as SNIA defines standard sets of data classification profiles across the different dimensions of IT services. These profiles will provide a blueprint for how to organize data classification SLOs and will align your organization with how ILM will evolve in the future.
Phase 3: This phase introduces the concept of "solution stacks." A solution stack conveys the idea of a set of homogenous data and storage services that have been instrumented to work together in an ILM context in support of an application and its information. A solution stack is a complete (or nearly complete) ILM environment around that application. For example, based on the SLOs for a specific application and its information, define where active primary storage goes; where long-term online retention is held; the protection, operational recovery, and disaster-recovery processes; and the regulatory and archive repositories. Begin operating this solution stack to ILM principles. This is the first step toward automating ILM around applications and their information requirements.
Phase 4: ILM automation will require the introduction of a new class of management tool that will automate IT infrastructure practices as homogenous islands. Why homogenous? It is far easier to first instrument a single vendor solution stack. Phase 4 implementations require that the supporting data and information services plus the network, server, and storage infrastructure be implemented so that they can be operated centrally. This requires standards and is one of the goals of the DMF s ILM plan.
Phase 5: Interoperability is achieved. Interoperability means IT products define their capabilities and provide instrumentation in terms of services provided and one product can be replaced with another in an IT infrastructure even if it provides those services in very different ways. Interoperability also means defining interfaces so one product can hand off services to another where it makes sense. Phase 5 scales from the largest of data centers down to branch offices so that a single set of processes, practices, and policies can be set in place and managed throughout the enterprise.
The value proposition of this vision for ILM is in its ability to solve the overwhelming complexity problem in the data center, achieving substantial reductions in total cost of ownership.
By being based on business requirements and standards, ILM is the perfect framework around which you can automate the data center.
The SNIA is engaged in the standards development efforts to make ILM a reality. To learn more about the association's efforts, or to make your voice heard, please visit www.snia.org/dmf.
Michael Peterson is program director of the SNIA's Data Management Forum, and Edgar St. Pierre is co-chair of the DMF ILM Technical Liaison Group and co-chair of the SNIA ILM Technical Work Group.