Abstract
Software virtualization in chemistry, mass spectrometry and cheminformatics is needed for software testing and development of software for different operating systems. In order to obtain maximum performance the virtualization software should be multi-core enabled and allow the use of multiprocessor configurations in the virtual machine environment. Server consolidation, by running multiple tasks and operating systems on a single physical machine, can lead to lower maintenance and hardware costs especially in small research labs. The use of virtual machines can prevent software virus infections and security breaches when used as a sandbox system for internet access and software testing. Complex software setups can be created with virtual machines and are easily deployed later to multiple computers for hands-on teaching classes. We discuss the popularity of bioinformatics compared to cheminformatics as well as the missing cheminformatics education at universities worldwide.
Introduction
Chemistry is a computationally fragmented field. A single research laboratory often requires Windows for instrument control (e.g., mass spectrometers), Linux for high-performance computing clusters, and macOS for data visualization and publication-quality graphics. Traditionally, this necessitated the maintenance of multiple physical machines, leading to high hardware costs and administrative overhead. Furthermore, chemistry software is frequently difficult to configure, and instrument-connected systems often run legacy operating systems like Windows XP to maintain driver compatibility, leaving them vulnerable to security threats.
Methods and Benchmarking
The study evaluated virtualization through three primary approaches:
- Performance Benchmarking: Comparing native Windows XP installations against virtualized "Guest" environments (Microsoft Virtual PC 2007). Tasks included isomer generation (Molgen), molecular descriptor calculation (CDK), and LC-MS data alignment (MZMine2).
- Server Consolidation: Utilizing the XEN Hypervisor to run multiple virtual servers (web, backup, and compute) on a single physical host.
- Educational Deployment: Creating a pre-configured virtual machine (VM) image for a mass spectrometry course, deployed across 41 classroom computers.
Results and Discussion
The research demonstrated that for single-core scientific tasks, the performance overhead of virtualization is remarkably low, typically between 5% and 10%. Benchmarks using NIST SciMark showed that VMs could achieve 91-100% of native speed.
As shown in Figure 3, the volume of research and citations in the field grew exponentially leading up to 2009. This growth underscores the need for scalable IT solutions. However, the authors noted a significant "Cheminformatics Gap." A Google API-based analysis of 325 university websites revealed that "bioinformatics" was mentioned nearly 100 times more frequently than "cheminformatics," suggesting a lag in formal education and infrastructure.
Virtualization offers a "sandbox" for security; if a VM is compromised by a virus or a software crash, it can be restored from a snapshot in seconds without affecting the host system. In educational settings, this allows for "instant" deployment of complex software environments, bypassing hours of manual configuration per student workstation.
One critical finding was the performance penalty for multi-threaded tasks. The study found a significant slowdown (approx. 50%) when using virtualization software that lacked Symmetric Multi-Processing (SMP) support, highlighting the necessity of multi-core enabled software like VMware or VirtualBox for modern scientific workloads.
Limitations and Future Outlook
Despite the benefits, hardware-level communication remains a challenge. Proprietary AD/DA cards used in many instruments often lack the 64-bit drivers required for modern host operating systems to pass through to a VM. Additionally, software licensing models in 2009 were not fully adapted to virtualized environments. The authors conclude that virtualization is a stepping stone toward cloud and grid computing, where hardware location becomes entirely irrelevant to the practicing chemist.
Reference
Kind, T., Leamy, T., Leary, J.A. et al. Software platform virtualization in chemistry research and university teaching. J Cheminform 1, 18 (2009).
Summary updated