Lorraine M. Herger

Design Alternatives for Virtual Interface Architecture and an Implementation on IBM Netfinity NT Cluster

The Virtual Interface Architecture (VIA) specification has been developed to standardize user-level network interfaces that provide low-latency, high-bandwidth communications. Few hardware and software implementations of VIA exist. Since the VIA specification is flexible, different choices exist for implementing various components of VIA such as doorbells, address translation methods, and completion queues. Although previous studies have evaluated the overall performance of different VIA implementations, there has not been a comparative study on the performance of VIA components. In this paper, we evaluate and compare the performance of different implementations of essential VIA components. We discuss the pros and cons of each design approach and describe the required support for implementing each of them. Then, we discuss an experimental implementation of the Virtual Interface Architecture for the IBM SP Switch-Connected NT cluster, one of the newest clustering platforms available. We discuss different design issues involved in this implementation. In particular, we explain how the virtual-to-physical address translation is implemented efficiently with a minimum Network Interface Card (NIC) memory requirement. We show how caching the VIA descriptors on the NIC can reduce the communication latency. We also present an efficient scheme for implementing the VIA doorbells without any hardware support. We provide a comprehensive performance evaluation study and discuss the impact of several hardware improvements on the performance of our implementation. The performance of the implemented VIA surpasses that of other existing software implementations of the VIA and is comparable to that of a hardware VIA implementation. The peak measured bandwidth for our system is 101.4MBytes/s and the one-way latency for short messages is 18.2?s.

Managing Software Assets in a Global Enterprise

The IBM Corp invents, develops and manufactures IT products, including computer systems and software, system networks, storage hardware and microelectronic products. The IBM CIO Global Infrastructure Architecture and Strategy (CIO GIAS) team is responsible for running and optimizing IBMs internal IT operations which include more than 500,000 users connecting to over 20,000 servers across 170 countries. The solution described below, implemented between 2008-2010, is a response to managing software licenses in one of the worlds largest distributed enterprises.

End-to-end service data analysis: Efficiencies achieved across the enterprise

Managing and optimizing a wide variety of information technology (IT) infrastructures, services, and data—ranging from prototypes and pilots to custom and production environments—in an efficient and cost-effective manner requires the ability to ingest and analyze a continuous stream of data from various sources. To complement this analysis, the IBM Research Integrated Solutions Group uses a high degree of automation for the ingestion and analysis of data. It is through these mechanisms that the overall intelligent framework we have developed is able to significantly advance the delivery of robust, secure, and compliant environments that perform at optimal levels for the worldwide IBM Research enterprise—which is distributed across 6 continents and 12 laboratories—as well as for the worldwide IBM corporate enterprise.

Exploit unstructured data using deep analytics to optimise enterprise IT asset management

Unstructured data pose a huge risk towards supporting critical goals of sustainable reconciliation of inventory due to over spending and audit exposure. They are inherently fragmented, incomplete and without restriction. Hence domain specific explicit information extracted from unstructured data may not be complete or contextually accurate or both for reconciliation. It requires inference to extract implicit information to make the extracted information usefully complete. Domain reconciliation modelled in relational database is static, requires deep domain knowledge prior to modelling and depends on surface commonalities of data labels and values between source and target. These design constraints make relational model-based reconciliation unfit for reconciling entities extracted from unstructured data. In this paper we propose an ontology-based semantic information model and semantic reconciliation mediator to extract valid entity information from unstructured data in knowledge format and reconcile them using pattern-based reconciliation. This agile, superior way of integrating information makes it possible to support unstructured information processing through inference and incorporates effects of temporal events that impact the ownership and usage rights of resources as well.

Statistics Versus Personal Experience: Stories of IBM Technical Working Women Around the World [The Good, the Bad, & the Ugly: Engineering FACTS]

For most women who enter the fields of science and engineering, the road is often circuitous, with detours along the way. The stories that accompany these career paths are not only interesting but provide guidance and insight to those who read them, who are perhaps searching for their own way forward. The goal of this article is to bring forward and illuminate the success stories of a diverse set of IBM technical women from around the world, demonstrating the rich and varied ways they have worked cross-culturally and globally to achieve professional success.

Leveraging cloud in the rapidly evolving workplace of the future

Rapid advances in cloud computing have made it possible to replace individual working environments with centralized and consolidated workplaces. These workplaces bring many advantages over traditional self-managed environments, such as improved security and compliance, and rapid on-boarding. In this paper we describe the implementation of a desktop cloud at IBM Research.