Techila Grid
 | |
| Developer(s) | Techila Technologies Ltd |
|---|---|
| Operating system | Windows, Linux |
| Type | distributed computing, middleware |
| License | Proprietary |
| Website | www.techilatechnologies.com |
Techila (earlier known as Techila Grid) is a commercial distributed computing middleware and management solution. It provides scalable performance to applications by enabling scalability across computing resources. The solution creates a computing service and execution environment, which is self-managing, based on autonomic computing. Techila is developed and sold by Techila Technologies Ltd, a privately held company headquartered in Tampere, Finland. According to IDC,[1] the solution enables organizations to create HPC infrastructure without the major capital investments and operating expenses required by new HPC hardware.
Product features
Techila is a distributed computing middleware and management solution, which can be used to access and manage IT resources for various high-performance computing (HPC) computing uses, including high-throughput computing (HTC) scenarios. The Techila technology is built on a patented autonomic computing architecture. The architecture has enabled features such as life-cycle management, self-configuration and self-healing, which simplify the deployment and operation of large-scale distributed computing systems. The basic idea of Techila is to create a scalable computing service and execution environment for applications with a workload, which can benefit of distributed computing. Common examples of workloads which are ready for distributed computing are embarrassingly parallel problems, such as parametric sweeps and simulations using Monte Carlo methods.
The Techila solution does not set limitations to the physical location of computing resources used. The Product Description of Techila lists three supported infrastructure designs:
- An infrastructure, which consists of on-premises resources.
- An infrastructure, which is built entirely in a cloud
- A hybrid infrastructure, which combines on-premises resources and resources from clouds.
The Techila technology started initially from the vision of grid computing and has evolved to a distributed computing middleware and management solution, which enables easy and secure utilization of available computing resources. Techila comes with productized support for many commonly used computing languages and environments. Techila has also an open application programming interface for applications developed by independent software vendors and systems Integrators. Techila offers also tools, which enable automated setup, integration, and management of cloud-based resources.
Techila was demonstrated by a research team at the University of Helsinki in 2011 as being capable of providing autonomic management to computing environments of large numbers of Windows Azure cloud instances. The University of Helsinki has also demonstrated Techila's ability to enhance the usability and utilization of large-scale cluster resources in projects implemented using MATLAB, R, Python, Java, and C/ C++/ C#.
In a Techila system, computational resources can be arranged into device groups for organizational, security, and administrative control. Despite its performance in large-scale systems, it is also suitable for smaller environments such as the TUTGrid which utilizes the idle capacity of desktop PCs and other computers at Tampere University of Technology (TUT) for scientific computing.
The Techila solution security was evaluated by Nixu Ltd in 2008. Nixu is Finland's largest specialist company in information security consulting, many global corporates as customers. After this, Techila has been accepted by security-sensitive industry sectors, such as Finance and Insurance, Manufacturing and Pharmaceutical.
Techila Server
Techila Server is a Java-based software product, which optimizes the Techila environment's performance and jobs in it. The optimization done by Techila Server supports not only large jobs, but also makes the system suitable for running small computational jobs. Techila's performance in different scenarios was evaluated in a thesis at Tampere University of Technology.[2]
Originally, the Techila Server was delivered as an embedded appliance. The embedded appliance product was discontinued in 2012. Currently, Techila Server is delivered either as a virtual appliance or using cloud-specific deployment tools.
Techila Worker
Techila Worker is the software agent middleware that must be installed on each computer that will participate in a Techila distributed computing environment. The computers can be physical or virtualized computers running on a hypervisor or in cloud. Techila supports following public cloud services: Microsoft Azure, Amazon EC2 (Amazon Elastic Compute Cloud) and Google Compute Engine. Once Techila Worker is installed on a computer, it will authenticate the computer to the Techila Server using a certificate, and automatically configures itself to run the jobs received from the Techila Server.
Techila Worker is a Java-based client middleware component, which can be run on Microsoft Windows or Linux. This enables client computers participating in the Techila system to be based on virtually any hardware platform or software platform. Techila Worker software runs on the lowest possible priority on the local computer. Techila Worker is also interoperable with batch-queuing systems, like the SLURM, TORQUE, or Oracle Grid Engine (previously known as Sun Grid Engine, SGE). This interoperability allows existing HPC users to use their existing infrastructures as a part of a hybrid cloud.
Administrator User Interface
A web-based Administrator User Interface provides administrators with a simplified and easy-to-use interface to the Techila Server. The Administrator User Interface allows monitoring system activity, view and control job execution, execution policy, monitor and control Workers and Worker groups, control security settings, and manage users.
Techila SDK
Techila SDK (earlier known as Techila Grid Management Kit or Techila GMK) is a software library installed on every End-User's computer. The End-User's computer can run any operating system with Java: Microsoft Windows, Linux, or Mac OS X.
Techila SDK provides application integration to the Techila environment. The SDK provides interfaces to many commonly used high performance computing application environments, such as MATLAB, R, Python, Perl, Java, .NET C/ C++ and FORTRAN.
External links
- Executive Summary of the Techila technology on Techila Technologies' web site
- Techila Product Description on Techila Technologies' web site
- Software Systems for Distributed Scientific Computing
- IDC Technology Spotlight - The Rise of Grid-Based High-Performance Computing: A Cost-Effective Approach to HPC Acquisition
- Evaluating the performance of job management systems in different distributed computing environments
- Innofinland Prizes of the President of the Republic 2009 to Techila
- ANSYS ADVANTAGE Volume VII - A Perfect Fit
- Techila Cloud Benchmark Report - Round 1
References
- ↑ Wu, Jie (2010). The Rise of Grid-Based High-Performance Computing: A Cost-Effective Approach to HPC Acquisition. IDC. p. 7.
- ↑ Koskinen, Marko (2010). Evaluating the performance of job management systems in different distributed computing environments (PDF). Tampere University of Technology. p. 63.