System Technologies and Services

Overview

The System Technologies and Services department at the Haifa Research Lab is one of IBM's technology innovation centers. Its mission is to develop leading-edge system and storage technologies for IBM's advanced IT products and services. The department is particularly strong in the areas of system software, system architecture, and system design and optimization tools, and supports IBM server and storage systems product lines and IBM Services business units.

We are in particularly interested in the emerging field of Cloud Computing, and we put a lot of effort on developing technologies to support this fascinating concept, as a part of the IBM Blue Cloud initiative. In particular, our work in system software focuses on the areas of virtualization and systems management, which we believe will be primary enabling technologies to make Cloud Computing become a marketplace reality. The department activities in this area influence a wide range of IBM system products, including IBM's xSeries systems (with a focus on blade servers), IBM's pSeries systems, IBM Systems Management, and more. We also develop advanced system and I/O architectures designed to deliver performance breakthroughs to IBM and the industry.

Our work in the area of storage systems focuses on developing new functions for data protection and availability, storage security, long-term data preservation, and storage power management as it relates to IBM's green technology initiative. A special focus area of our team is storage support for Cloud Computing. In this project, we investigate the architectural challenges of designing a highly scalable and cost-effective storage infrastructure, allowing to easily store and access data objects, and mobilize them across the Cloud.

A large part of our team conducts research in the areas of improving system performance and quality, specifically targeting the special challenges presented by multicore architectures. Our work in this area includes tools and methods for designing reliable and correct systems, and for improving software performance using powerful compiler and post-link optimizations. Our work in these areas has resulted in significant performance improvements and reliability enhancements to many important applications and systems of IBM, customers and partners, and has been recognized by many papers and patents.

Motivated by the proliferation of IT services as an emerging marketplace trend, we have recently started to conduct research on optimizing IT service delivery - an area which is receiving noticable attention from both industry and academia. For example, we work closely with IBM field professionals in the area of Business Continuity and Resiliency Services, where we contribute by developing capabilities for information protection, disaster recovery and more.

During the recent years, our team has achieved several landmarks in computer system technologies state-of-the-art. Our team has played a key role in the development of IBM's Virtualization Manager Software, announced in late 2006. This technology was recently recognized as the winner of the ServerWatch Product Excellence Award for 2007, and honored with the prestigious IBM Research Division Accomplishment Award for 2007. The compiler and post-link optimizations developed by our team have helped IBM Power systems become the leader in TPC-C,SPEC-CPU and SAP benchmarks for Linux, and are used to competitively optimize IBM systems for commercial workloads. Some other well-known examples of work done in our department are the invention of the now standard iSCSI protocol and Object Storage technology; contributions to IBM's Blue Gene supercomputer; and our work on the advanced Copy Services in IBM's flagship storage platform. For these and other achievements, our team has been recognized multiple times by the IBM Research Division Accomplishment Award and received other honors, such as recognition of some of our team members as 2003 InfoWorld Innovators.


Collaborations and Partnerships

Across all our activities, our team works with internal IBM product organizations as well as with IBM customers. We engage directly in development and delivery projects, partnering with the following IBM business units:

We routinely work with standard organizations to set systems and storage industry standards. A recent example of our work in this area is the CBCS standard for secure access to networked storage. Our team has conceived and developed this standard, and then led the process of its approval by the T10 standards body.

Lastly, we collaborate with academic and industrial organizations on projects of mutual joint interest. These include the following activities:

  • Reservoir. This is an ambitious EU FP7 project, where we lead a consortium of 13 European companies and universities who collaborate on developing an advanced infrastructure for Cloud Computing, based on deep integration of virtualization and grid technologies.
  • Nexof-RA. This is an EU FP7 project where we participate in the development of a Reference Architecture for the NESSI open service framework. Our contribution to this project focuses on IT infrastructure aspects, and is based on the Reservoir architecture.
  • Shadows. This is an EU FP6 project, led by our team, which focuses on developing tools and methods for designing self-managing systems, based on the DMTF/CIM and Eclipse/TPTP open standards.
  • CASPAR. This is an EU FP6 project which focuses on tools and methods for preserving cultural, scientific, and artistic long-term data. The project builds on the Object Storage standard (conceived by our team and standardized by SNIA) and the OAIS standard.
  • HiPEAC. This is an European Network of Excellence on High Performance and Embedded Architecture and compilation. HIPEAC started as an FP6 project, and now continues as an FP7 project.
  • SARC. This is an FP6 Integrated Project in the area of advanced computer architectures, with a focus on a systematic scalable approach to the design of systems ranging from small embedded systems to large scale networked data servers.
  • ACOTES. This is an FP7 project aimed at addressing the complexity and cost of programming emerging embedded architectures by implementing program transformation techniques in compilation tools to enable efficient programming of highly parallel architectures for stream computing.
  • MilePost. This is an FP7 project that exploits machine learning for embedded programs optimizations. The overall objective of MilePost is to develop compiler technology that can automatically learn how to best optimize programs.

Departments

Selected Research Papers - System and Storage Platforms

  • "Virtual machine time travel using continuous data protection and checkpointing", Paula Ta-Shma, Guy Laden, Muli Ben-Yehuda, and Michael Factor, ACM SIGOPS Operating Systems Review, 42:1 (2008).
  • "Tapping into the fountain of CPUs: on operating system support for programmable devices", Yaron Weinsberg, Danny Dolev, Tal Anker, Muli Ben-Yehuda, Pete Wyckoff, Proceedings of ASPLOS 2008: 179-188.
  • "Multitoroidal interconnects for tightly coupled supercomputers", Yariv Aridor, Tamar Domany, Oleg Goldshmidt, Yevgeny Kliteynik, Edi Shmueli and Jose Moreira, IEEE Transactions on Parallel and Distributed Systems 19(1):52-65 (2008).
  • "B-trees, shadowing, and clones", Ohad Rodeh, ACM Transactions on Storage 3(4): (2008).
  • "Preservation DataStores: New storage paradigm for preservation environments", Simona Rabinovici-Cohen, Michael Factor, Dalit Naor, Leeat Ramati, Petra Reshef, Shahar Ronen, Julian Satran, David Giaretta, IBM Journal of Research and Development, Special Issue on Storage Technologies and Systems, 2008.
  • "The ANSI T10 object-based storage standard and current implementations", David Nagle, Michael Factor, Sami Iren, Dalit Naor,Erik Riedel, Ohad Rodeh, Julian Satran, IBM Journal of Research and Development, Special Issue on Storage Technologies and Systems, 2008.
  • "Architectures for Controller Based CDP", Guy Laden, Paula Ta-Shma, Eitan Yaffe, Michael Factor, and Shachar Fienblit, Proceedings of The 5th USENIX Conference on File and Storage Technologies, February 2007 (FAST '07).
  • "Capability based Secure Access Control to Networked Storage Devices", Michael Factor, Dalit Naor, Eran Rom, Julian Satran, Sivan Tal, Proceedings of MSST 2007:114-128.
  • "Preservation DataStores: Architecture for Preservation Aware Storage", Michael Factor, Dalit Naor, Simona Rabinovici-Cohen, Leeat Ramati, Petra Reshef, Julian Satran, David L. Giaretta, Proceedings of MSST 2007: 3-15.
  • "A Self-optimized Job Scheduler for Heterogeneous Server Clusters", Elad Yom-Tov, Yariv Aridor, Proceedings of JSSPP 2007:169-187.
  • "The need for preservation aware storage: a position paper", Michael Factor, Dalit Naor, Simona Rabinovici-Cohen, Leeat Ramati, Petra Reshef, Julian Satran, Operating Systems Review 41(1): 19-23 (2007).
  • "Encompass: Managing Functionality", Oleg Goldshmidt, Benny Rochwerger, Alex Glikson, Inbar Shapira, Tamar Domany, Proceedings of IPDPS 2007:1-5.
  • "PANACEA Towards a Self-healing Development Framework", David Breitgand, Maayan Goldstein, Ealan Henis, Onn Shehory, Yaron Weinsberg, Proceedings of Integrated Network Management 2007:169-178.
  • "On Fully Distributed Adaptive Load Balancing", David Breitgand, Rami Cohen, Amir Nahir, Danny Raz, Proceedings of DSOM 2007: 74-85.
  • "Using the Right Amount of Monitoring in Adaptive Load Sharing", David Breitgand, Rami Cohen, Amir Nahir, Danny Raz, Proceedings of ICAC 2007: 7.
  • "Cost Aware Adaptive Load Sharing", David Breitgand, Rami Cohen, Amir Nahir, Danny Raz, Proceedings of IWSOS 2007: 208-224.
  • "IP Only Server", Muli Ben-Yehuda, Oleg Goldshmidt, Elliot K. Kolodner, Zorik Machulsky, Vadim Makhervaks, Julian Satran, Marc Segal, Leah Shalev, Ilan Shimony, USENIX Annual Technical Conference, General Track 2006: 381-386.
  • "Automated and Adaptive Threshold Setting: Enabling Technology for Autonomy and Self-Management", David Breitgand, Ealan Henis, and Onn Shehory, The 2nd IEEE International Conference on Autonomic Computing (ICAC-05), 2005.
  • "Design of the iSCSI Protocol", Kalman Z. Meth, Julian Satran, IEEE Symposium on Mass Storage Systems 2003:116-122.
  • "Internet Protocol storage area networks", Prasenjit Sarkar, Kaladhar Voruganti, Kalman Z. Meth, Ofer Biran, Julian Satran, IBM Systems Journal 42(2): 218-231 (2003)
  • "Management of application complexes in multitier clustered systems", Antonio Abbondanzio, Yariv Aridor, Ofer Biran, Liana Fong, German S. Goldszmidt, Richard Harper, Srirama Krishnakumar, Gregory Pruett, Ben-Ami Yassur: IBM Systems Journal 42(1), 2003.

Selected Research Papers - Code Optimization and Quality Technologies


For a full list of publications, see the department publications page: http://w3.haifa.ibm.com/dept/svt/code_public.html

Senior Manager

Yaron Wolfsthal,