Strategy Areas

The focus of our research in advanced storage controllers is to develop technologies that enhance the integrity, reliability, scalability, energy efficiency, performance and cost-effectiveness of storage systems and controllers. Some of our current research areas include energy proportionality, data protection and quality of service. Our group has made several contributions to IBM storage products including the flagship enterprise storage system---IBM DS8000, and the virtualization appliance---TotalStorage SAN Volume Controller. Additional recent contributions to IBM products include innovations in caching and prefetching, such as the Adaptive Replacement Cache (ARC), Wise Ordering for Writes (WOW), and Adaptive Multi-stream Prefetching (AMP) algorithms, and data reliability technologies.

Projects

Archive systems research includes a wide range of projects related to long-term storage, archiving, and data preservation, and the technologies that enable them. Our group includes recognized experts in data compression and deduplication technologies as well as expertise in many disciplines that are crucial to efficient and reliable data storage and protection. One of our largest ongoing projects is the IBM Information Archive, a scalable, policy-managed archive appliance that will be released as an IBM product in 2010. The group is also involved in several efforts around tape technology to make it a more attractive and viable vehicle for data archive and interchange; these include the LTFS file system for LTO Generation 5 tape, and a project aimed at a new type of tape storage that is higher-capacity and better suited to long-term archive than existing tape offerings. Additionally, the group has a long history in the the research and enhancement of Tivoli Storage Manager (TSM).

Projects

Explosive growth in computing and storage requirements has made enterprise systems management increasingly complex for administrators. Supporting efficient and reliable access to data and at the same time reducing operating cost and total cost of ownership (TCO) is extremely difficult. The Autonomic Systems and Storage Management group at IBM Research Almaden, is conducting research on reducing cost of managing next generation storage and compute cloud environments through simplification and automation. This includes research in building innovative and intelligent user interface design for systems management to low level planning, analysis and optimization of underlying compute and storage resources through technologies such as virtualization. Our research agenda consists of both short and long term goals and we partner with other research and product groups within IBM to solve practical and challenging problems faced by our partners and customers.

Some of our current focus areas include:

• Systems management interface to analyze topology, dependencies and configuration of different cloud components
• Integrated management of server and storage virtualization
• Scalable and efficient configuration discovery and monitoring
• Integrated and end-to-end planning, provisioning and optimization in storage cloud
• Power efficient management of cloud resources and workload
• Change tracking and configuration history management
• Chargeback modeling and monitoring for clouds
• Real-time and low cost performance analysis
• Replication and disaster recovery planning and management

Projects

The File Systems team explores and develops new technologies in file systems and facilitates using these technologies in IBM products.

The File Systems team at IBM Research Almaden originated the General Parallel File System (GPFS), IBM's parallel, shared-disk file system for cluster computers. It is available on the IBM e(logo)server. pSeries. and on Linux clusters. GPFS is used on many of the largest supercomputers in the world and is also used in commercial applications such as database, file serving, digital media and content management. Almaden researchers play an ongoing role in the evolution and deployment of new GPFS releases.

We are also working on File System Federation, which is a standards-based method of replicating and migrating data among multiple NFS V4 servers, and Scale-out File Serving, which is software built in IBM's cluster file systems to provide a gateway solution to broaden the range of environments that can benefit from storage consolidation and advanced data management.

Projects

Solid state storage represents an exciting and challenging area of research. We are looking into incorporating solid state technology into storage systems in several ways. Solid state disks (SSDs) based on flash memory present opportunities to dramatically increase the performance of storage systems; however, there are challenges to taking advantage of the full potential of SSDs such as cost, architectural considerations and more. We are developing Easy Tier, a technology that maximizes the value of SSDs when incorporated into storage systems. We are also exploring architectural approaches to storage systems that can utilize flash memory effectively, and are colloaborating with others to look at system implications of newer technologies such as phase change memory.

Projects

We are designing a highly distributed storage cloud with the capability for Smart Data, data that moves from producer to consumer seamlessly without user intervention and analyzes network bandwidth, storage capacity and device capabilities, including smart phone to server. The unique IBM Research differentiating technologies are:

• Panache : Globally distributed scalable peer-to-peer data movement technology with support for caching and high availability
• Cloud Content Store (CCS) : Secure cloud storage for unstructured data and content depots that is continuously available everywhere for any device and application
• Hadoop-on-GPFS : Business and Infrastructure Analytics running real time on the large swaths of data in the cloud
• Leopard and Dynamo : Scalable storage management infrastructure to deliver the cloud service at costs below those of our competitors

Storage clouds will become smarter in terms of data movement and application analytics and will have the ability to host many applications such as customer sentiment analysis, Smart Grid analysis, financial services applications and more.

Projects