Storage Systems - Projects - Storage Management

Overview

Sandeep Gopisetty, Manager, Senior Technical Staff Member, Autonomic Storage Management Our vision is to develop integrated storage network analytics and performance management tools to aid policy-based autonomic resource provisioning of storage networks. These technologies will help system administrators to seamlessly manage the various aspects of storage networking: monitoring, performance, planning and provisioning in heterogeneous storage environments. Our vision led to the creation of TotalStorage Productivity Center, an enterprise storage management product, focussing on advanced functions -- such as self-diagnosis, error prediction, configuration checker, policy based storage management, SAN configuration and planning, application migration and disaster-recovery planning.

TotalStorage Productivity Center

In partnership with Tivoli and STG, Almaden Storage Systems researchers are making a positive impression with customers and adding to IBM's bottom line by architecting, designing and implementing Tivoli's TotalStorage Productivity Center product.

Within just three quarters of unveiling TPC Version 3.1, the number of customers has quadrupled. TPC 3.1 greatly improves and simplifies the processes customers use to manage the storage aspects of their rapidly growing volume of business-critical information, applications and tools. This is achieved using open-system autonomic policy-based protocols to manage heterogeneous storage infrastructures from a single point of control.

While previous TPC versions had four different graphical user interfaces (GUIs) and four different databases, now there is only one, which is seamlessly integrated. Installation time was reduced from days to a few hours, and the number of installation CDs down from more than 300 to just 20, which support 14 languages. Most impressive was the new "topology viewer" user interface, which provides a central location and graphical view of the storage environment, which enables users to monitor and troubleshoot problems quickly, and to access additional TPC tasks and functions without the users losing their orientations to the overall environment. Some of the features include:

• Progressive information disclosure reduces the complexity of system management by unveiling complexity only as the user requests or requires more details or options regarding particular aspects of the system
  
• Enriching the environment with overlays that indicate such qualities as device health, performance and alerts
  
• Semantic Zooming: A new type of context-aware graphical zoom that keeps the user-environment simple and focused on the task at hand – reducing unnecessary details -- even when used with extremely large and complex networks. Scaling is an issue in all kinds of complex displays
• Pinning allows users to choose VIP items that will always be kept in view.
  
• Dynamic Grouping allows users to rearrange their views to focus on portions of their environment needed for their current tasks.
  

The new features include:

• Policy-based SAN configuration validation, using best practices
• Graphical datapath topology, which will aid in locating problems, provides spatial view and helps in isolating faults and identifying data bottlenecks
• A temporal view of history of configuration changes, including a "time-travel" viewer for examining previous configurations

The new integrated planner (host, security, storage) can provision storage in minutes (rather than several weeks). It allows for multipath configuration, supporting failover, load balancing and round robin. It also helps configure the fabric using autonomic policy-based zoning and provides advice and configuration of storage volumes based on workload performance profiles.

CIM Client

The SNIA (Storage Networking Industry Association) SMI-S (Storage Management Initiative Specification) standard is a significant and relatively recent evolution in storage management. SMI-S enables centralized management of distributed storage assets, and is the foundation for our Total Storage Productivity Center (TPC) product. The SMI-S standard is implemented using the CIM (Common Information Model) DMTF (Distributed Management Task Force) industry-standard management architecture. A CIM Client is a client that issues CIM Operation Requests, CIM Message Requests and receives and processes CIM Operation Responses. The CIM Client was open-sourced through the Standards Based Linux Instrumentation for Manageability (SBLIM) project within the Linux Technology Center.

APERI

Vendor-neutral, open, storage management framework designed to cultivate both an open-source community and an ecosystem for complementary products, capabilities, and services around the framework to promote greater consumer choice and foster competition.

• Provide an open, extensible, standards-based storage management framework
• Give customers more flexibility and choice on how to manage their storage
• Simplify the infrastructure customers need to manage storage

SAN Simulator

Large storage installations are extremely complex today; as storage capacities grow, complexity increases. One of the primary constraints that developers face is the inability to test any SAN software against large, complex SAN environments, primarily due to lack of hardware resources. It is a serious waste to own expensive SAN boxes just for testing software. Our SAN Simulator will simulate a SAN environment through software. It will allow the user to create a SAN configuration, add devices to the SAN, create arbitrary connections between devices and remove connections between devices. Since most devices today expose their services through a Common Information Model Object Manager (CIMOM), the simulator will create CIM instances based on user specifications. The simulator will generate various profiles, such as, fabric, array and disk profiles.

Atlantis

The field of disaster recovery (DR) has gained a lot of prominence after the 9/11 terrorist attacks. Businesses now want redundancy in people, processes and information technology (IT) infrastructure so they can withstand machine or site failures, virus attacks, city-wide shutdowns etc. To address disaster recovery at the IT level, one has to provide redundancy at a combination of application, server, network and storage levels. Furthermore, redundancy at the storage level can be provided at a combination of database, file-system or block-storage levels. The goal of our Atlantis project is to provide automated disaster recovery planning at the storage level by combining the diverse storage DR planning technologies and best practices into an integrated framework.

Dakshi Agrawal, James Giles, Kang-Won Lee, Kaladhar Voruganti and Khalid Adib-Filali, "Policy-based Validation of SAN configuration", in IEEE Policy 2004.

Kaladhar Voruganti and Sandeep Gopisetty, "Enterprise Storage Management", in International Workshop on IT-enabled Manufacturing, Logistics and Supply-Chain Management", 2003.

Kostas Magoutis, Murthy Devarakonda, Norbert Vogl,Kaladhar Voruganti, "RAIC: Scaling Storage Availability and Performance in On-Demand Data Centers", in 1st Workshop on Operating System and Architectural Support for the on demand IT Infrastructure, OASIS, 2004.

Kaladhar Voruganti, Jai Menon, Sandeep Gopisetty, "Land Below a DBMS", in ACM SIGMOD Record, March, 2004.

Aameek Singh, Kaladhar Voruganti, Sandeep Gopisetty, David Pease, et. al, "A Hybrid Access Model for Storage Area Networks", in IEEE NASA MSST, 2005.

Aameek Singh, Kaladhar Voruganti, Sandeep Gopisetty, David Pease, et. al, "Security vs Performance: Tradeoffs using a Trust Framework", in IEEE NASA MSST, 2005.

Aameek Singh, Kaladhar Voruganti, Sandeep Gopisetty, Aki Fleshler, Ramani Routray, Chung-hao Tan, "SANFS Maestro: Resource Planning for Enterprise Storage Area Network FileSystems", International Conference on e-Business, e-government and outsourcing, EEE, 2005.

Aameek Singh, Madhukar Korupolu, Kaladhar Voruganti, "Zodiac: Efficient Impact Analysis for SANs", in USENIX FAST 2005.

Sandeep Uttamchandani, Kaladhar Voruganti, David Pease, et. al, "Polus: Growing Storage QoS management beyond a 4-year old kid", in USENIX FAST 2004.

Prasenjit Sarkar, "Instant Image: Transitive and Cyclical Snapshots in Distributed Storage Volumes", Lecture Notes in Computer Science: Proceedings of Europar 2000. August 2000.