VMware Academic Program
Committed to strengthening VMware’s relationship with the academic and research communities.

VMware Academic Program – Research Projects

VMware Academic Program – Research Projects

VMAP Collaborations

Academic_Program_SigBanner-FINAL

The VMware Academic Program (VMAP) supports a number of academic research projects across
a range of technical areas. We initiate an annual Request for Proposals (RFP), and also support a
small number of additional projects that address particular areas of interest to VMware.

Each year VMAP sponsors a Request for Proposals (RFP). A couple of past topics have been Security for Virtualized and Cloud Platforms and 2014,  Datacenter automation in Virtualized Environments.

VMAP has several industrial affiliations with schools both herein the U. S. and abroad.

Current Collaborations

Affiliations

Past Collaborations

Georgia Institute of Technology Carnegie Mellon University Boston University
Harvey Mudd College Georgia Institute of Technology Carnegie Mellon University
Rutgers University  Massachusetts Institute of Technology Imperial College London
University of California, Riverside Stanford University Israel Institute of Technology-Technion
University of Missouri University of California, Berkeley Karlsruhe Institute of Technology
University of North Carolina, Chapel Hill University of California, San Diego Massachusetts Institute of Technology
University of Texas, Dallas University of Wisconsin New York University
University of Wurzburg University of Tennessee
University of Toronto

Current Collaborations


Top

georgia-tech-logo kishore

Academic Lead: Umakishore Ramachandran

Rethinking Operating system Structure with Heterogeneous Main Memory

For several decades, DRAM has been used as the main memory of a CPU due to its low latency and long endurance.  However, considerations of cost, power consumption, and heat dissipation limit the amount of DRAM that can be provisioned for a single CPU. These considerations are even more severe in a datacenter used for giant-scale services both due to the large number of server machines involved and the large datasets of applications running on such server farms.  Flash Memory, a non-volatile memory technology, has been a great success in the past decade.  Flash memory is denser, cheaper, and more power-efficient than DRAM. It is used in most mobile devices such as smartphones and tablets. In addition, it is finding increasing adoption in enterprise class infrastructure as well. However, since its access granularity and latency properties are more akin to that of the hard disk, thus far it has been typically used as a replacement for the hard disk and not for the main memory.  Phase Change Memory (PCM) is a new non-volatile memory technology that is on the horizon for mass production. Different from Flash, it is byte addressable, and its latency properties are closer to that of the DRAM. Moreover, it has better endurance properties than Flash.  In this proposal, we would like to explore an alternate model for the memory system architecture of a compute node in a datacenter and its implication for system software. In this model all three memory technologies appear at the same level so far as the CPU is concerned. While the DRAM is volatile, the other two are not. Each technology has its unique characteristics in terms of access granularity, access times, lifetime, and cost per bit. Therefore we can use each of them in a different role from the point of view of system software. Using each technology for a different role could lead to a revolutionary approach to the design of the system software. Our proposed research comprises two parts: (a) rethink the subsystems of the entire system software stack such as the virtual memory manager, the file system, the scheduler, the programming interfaces, and the execution model in light of this structure; and (b) rethink the memory architecture in support of the system software stack.

Top

harvey-mudd kuenning

Academic Lead:Geoff Kuenning

VMware Probes Project

The Computer Science Clinic consists of a small group of undergraduate looking into making vProbes accessible to non-programmers by exposing some of its functionality via a web-based user-interface. vProbes is a flexible dynamic instrumentation tool developed at VMware to help with understanding what is occurring in a virtual machine or on the virtual infrastructure on which the Virtual Machine runs.  For more details about this project, a summary is available.

Top

rutgers-logo abhishek

Academic Lead: Abhishek Bhattacharjee

Virtualization for Systems with Emerging Storage and Memory Technologies

Sever systems are currently undergoing a radical change in design options for storage technologies and the memory hierarchy. On the storage side, alternatives to traditional hard-disk drives (HDDs) like Solid State Drives (SSDs), phase-change memories (PCMs), and various hybrid approaches are becoming commonplace. Meanwhile, memory technologies are rapidly changing, including innovations like 3D die stacking, which integrates multiple disparate silicon die via high bandwidth and low-latency interconnects. While 3D stacking can realize logic or SRAM caches, initial proposals use them to implement DRAM caches (accessed after L1, L2, and L3 caches).

These changes in storage and main memory require appropriate hardware/software support for virtual memory. This proposal considers this support, focusing on Memory Management Unit design, page table design, and the role of large pages, particularly in the context of virtualization. As virtual machines are expected to run on server systems with these innovations, our observations and design insights will be crucial for future computer systems.

Top

ucriverside harsha

Academic Lead: Harsha Madhyastha

Radio+Tuner: Tunable Storage for Software defined Datacenters

We propose to develop the Radio object store 1) that is amenable to be configured differently to meet theneeds of different workloads, and 2) whose implementation can be incrementally evolved over time as new workloads emerge. The key feature that will distinguish Radio from other storage systems is that it will make no a priori assumptions about its input workload or the associated performance goals. Instead, our approach is to separate mechanism (“how is the system implemented?”) from policy (“what configuration is deployed?”).

Top

unc reiter

Academic Lead:Michael Reiter

Timing Side-Channels in Modern Cloud Environments

Prof. Reiter proposes to develop privacy preserving techniques to support applications that utilize pattern matching algorithms as a major component, with special focus on cloud scenarios where a data owner outsources data to the cloud to enable secure and authorized sharing of information with third-party “clients”. Pattern matching is fundamental to numerous popular applications, such as information retrieval, analysis of genomic data, and malware detection, to name a few. Prof. Reiter proposes to develop techniques that can be applied to this wide range of applications while minimizing the trust placed in the cloud and third-party clients. Such techniques will not only benefit data owners and third parties by better insuring their privacy but will also benefit clouds and third-party service providers by reducing their liability to secure users’ data.

Top

missouri calyam

Academic Lead: Prasad Calyam

Title: VMLab: Infrastructure to Support Desktop Virtualization Experiments for Research and Education

In the VMLab project, Dr. Calyam and his research group at University of Missouri-Columbia are developing a novel community infrastructure for supporting desktop virtualization experiments for research and education use cases. Their project activities include: (a) conducting research and development relating to virtual desktop cloud (VDC) resource allocation and thin-client performance benchmarking, (b) providing scalable access to desktop virtualization sandboxes for system administrators and educators, (c) provisioning virtual desktops for classroom lab user trials involving faculty and students within a federated university system environment, (d) evaluation of the feasibility to deploy computationally intensive ‘interactive applications’ (e.g., remote volume visualization) and ‘simulation-as-a-service’ in virtual desktops, and (e) educational laboratory course curriculum development involving desktop virtualization exercises.

Top

utdallaslin

Academic Lead:Zhiqiang Lin

VULCAN: Automatically Generating Tools for Virtual Machine Introspection Using Legacy Binary Code

This project will design and develop an advanced system, VULCAN, to automatically generate the VMI tools. It builds upon a system called VM Space Traveler (VMST), that bridges the semantic gaps using an online kernel data redirection approach. While VMST has laid a technical foundation for VULCAN especially how to bridge the semantic-gap, it only works for Linux kernel with limited capabilities. In this project, VULCAN will extend VMST in non-trivial ways. In particular, VULCAN aims to automatically generate VMI tools for: (1) memory introspection, (2) disk data introspection, (3) Linux kernels (with a series of different kernel versions), and (4) Microsoft Windows kernels.

Top

wurzburg    kounev

 

 

 

Academic Lead:  Samuel Kounev

Online Model-based Performance and Resource Management in Virtualized Application Environments

Modern IT systems are increasingly deployed on virtualized infrastructures for maximum flexibility and cost efficiency. To reduce operating costs and improve energy efficiency, resource allocations need to be continuously optimized while ensuring that the hosted IT services meet their SLAs. Unfortunately, existing tools for performance and resource management of virtualized infrastructures lack the ability to proactively determine the effects of changing resource allocations on the performance of hosted IT services. The aim of the project is to develop a novel framework for online model-based performance and resource management, and integrate this framework in VMware vSphere. The framework will use mathematical performance models to predict the performance of applications hosted in the virtualized environment for a given workload and resource allocation. The models will be inferred automatically during operation and will be analyzed on-the-fly to predict the effects of dynamic changes in the environment (e.g., changing workloads or system reconfigurations). The online prediction will be leveraged to continuously optimize resource allocations maximizing resource efficiency while ensuring SLA compliance. The modeling approach used in this project is based on the Descartes Meta-Model (DMM) – a new architecture-level modeling language for modeling quality-of-service and resource management related aspects of modern dynamic IT systems, infrastructures and services.

Top


Affiliations


cmu-logo pdl

Parallel Data Laboratory

Top

georgia-tech-logo cercs

Center for Experimental Research in Computer Systems

Top

mit-logo

Industrial Liaison Program

Top

stanford sedcl

Stanford Experimental Data Center Laboratory

Top

berkeleyamp-lab

Algorithms Machines People Lab

 

Top

uc-sandiego

Top

wisconsin

WISDoM

Top


Past Collaborations

bu-logo coskun1-75x75

Academic Lead: Ayse Coskun

Energy-Efficient and Reliable Server Consolidation for the Cloud

Datacenter energy consumption reaches 3% of total US electricity use today and increases by 15% every year, costing billions of dollars. Close to half of this cost is for cooling. In addition, high power consumption increases temperature, and high temperatures accelerate reliability degradation. A significant portion of the hardware failures during useful life of chips are currently caused by high temperatures. Such failures do not only cause customer dissatisfaction but also delay delivering results because of the need of rerunning the requested tasks and increase the overall energy use.

cmu-logo ganger

Academic Lead: Greg Ganger

Vcloud at Carnegie Mellon: Establishment, Automation, Measurement, and Evolution

In a large-scale collaboration, we are standing up and operating a state-of-the-art cloud based on VMware software and HP hardware (with Intel processors and Samsung DRAM and SSDs) implementing the vCloud APIs. The PDL vCloud will replace a multitude of single-purpose clusters, managed and underutilized by individual groups, with an IaaS private cloud for class projects, simulations, data analyses, and cluster and data-intensive computing activities. Moreover, it is an invaluable resource for studying the usage patterns and demands placed on clouds in academic settings, whose workloads we think will be representative of analytics-heavy non-academic activities. We are deeply instrumenting the infrastructure, at various levels and in cooperation with VMware and HP experts, working with different groups to understand their specific requirements, and exploring/developing automation approaches tuned to the challenging characteristics of not-for-profit academic environments.

Top

imperial-college yoshida

Academic Lead: Nobuko Yoshida

A Robust Framework of conversational Assurance for Distributed Systems Based on Multiparty Session Types

The aim of this collaboration is to support a PhD student whose project is to enable safety assurance at the level of system-level messages, as given in messaging architecture (message-oriented middleware) through the establishment of supporting theories and architectural research and development. The assurance is to be done against high-level specifications, given for applications programs and distributed infrastructure.

Top

technion

Managing Hierarchical Data in MySQL

    In this project the students will supply a tool to insert environments to mySQL database using the nested set model. When a new kind of environment should be insert to the database the tool must provide an interface (web preferably) to mention the new environment parent and childes. The tool alters the representation inside the database according to the nested set model.

Top

kit_logo_V2 kounev

Academic Lead: Samuel Kounev

Online Model-based Performance and Resource Management in Virtualized Application Environments

Modern IT systems are increasingly deployed on virtualized infrastructures for maximum flexibility and cost efficiency. To reduce operating costs and improve energy efficiency, resource allocations need to be continuously optimized while ensuring that the hosted IT services meet their SLAs. Unfortunately, existing tools for performance and resource management of virtualized infrastructures lack the ability to proactively determine the effects of changing resource allocations on the performance of hosted IT services. The aim of the project is to develop a novel framework for online model-based performance and resource management, and integrate this framework in VMware vSphere. The framework will use mathematical performance models to predict the performance of applications hosted in the virtualized environment for a given workload and resource allocation. The models will be inferred automatically during operation and will be analyzed on-the-fly to predict the effects of dynamic changes in the environment (e.g., changing workloads or system reconfigurations). The online prediction will be leveraged to continuously optimize resource allocations maximizing resource efficiency while ensuring SLA compliance. The modeling approach used in this project is based on the Descartes Meta-Model (DMM) – a new architecture-level modeling language for modeling quality-of-service and resource management related aspects of modern dynamic IT systems, infrastructures and services.

Top


mit-logo

1. Alan Edelman, High Performance computing 2: Flexible, Elastic, and Convenient

2.  Una-May O’Reilly, Machine Learning Approach Supporting VM Research Management which Explicity Fulfills Service

3.  Una-May O’Reilly, Machine Learning to Optimize Resource Allocation from a Service Level Agreement Perspective

edelman

Academic Lead: Alan Edelman

High Performance Computing2: Flexible, Elastic, and Convenient

oreilly

Academic Lead: Una-May O’Reilly

Machine Learning Approach Supporting VM resource Management which Explicitly Fulfills Service Level Agreements

Machine Learning to Optimize Resource Allocation from a Service Level Agreement Perspective

Top

nyu-logo dodis

Academic Lead: Yevgeniy Dodis

Random Number Generation in Virtualized Environments

As computing is rushing headlong into an era of ubiquitous virtualization because of applications like cloud computing, desktop security, and others, functionality and improved performance have been a core target of research and development regarding virtualization. In contrast, in this project we consider instead the new *security problems being introduced* by the move to virtualized settings. In particular, we will concentrate on designing secure random number generators (RNGs) in vitualized environments. Recall, RNG is a computing process whose goal is to provide private, uniformly distributed bits to applications. Designing cryptographically strong RNGs is a major challenge even in traditional settings, but many new vulnerabilities arise when such RNGs collide with virtualization. The proposed research will respond by setting the theoretical and architectural foundations for secure RNG design and use in virtualized settings, as well as improve various aspects of such RNG designs even in traditional, non-virtualized environements.

UTWorkmarksALL terpstra

Academic Lead: Dan Terpstra

Hardware Performance Monitoring of Scientific Applications in VMware Environments

Top

utoronto delara

Academic Lead: Eyal de Lara

Flexible Computing with VM Fork

Top

Leave a Reply

Your email address will not be published. Required fields are marked *