Posted by Tracy Valentine
As mentioned in our previous post on virtualization, the advent of client-server topology thankfully put computing power into the hands of end users, but as these applications were rolled out over many years a number of financial and technical issues became readily apparent.
- Low Infrastructure Utilization – often times operating at less than 25% of total capacity
- Increased Infrastructure Cost – for servers, desktop computers, facilities and cooling
- Increased Maintenance Cost – more network administrators and server upgrades
- Insufficient Disaster Protection – complex networks became more difficult to backup
Coming to the rescue was the implementation of virtualization technology inside x86-based hardware. Pioneered by VMware, this new layer of software allowed multiple “virtual machines” to operate on a single server. Taking advantage of multi-core, multi-processor hardware, these virtual machines act like a single physical computer, complete with CPU, memory, hard drive storage, and network interface.
Operating systems, which are loaded onto each of the virtual machines, behave as though they were running on a dedicated server, even though this fully isolated environment is software based. These operating systems, which can include Microsoft Windows running alongside various flavors of Linux, are completely isolated from each other. If one virtual machine encounters a problem and crashes, the others are unaffected while the problem is resolved and the one machine is rebooted separately.
Beyond the level of a single server, users can now create what VMware calls a Virtual Infrastructure which lets the IT group manage a series of servers as though they were a single machine. In this case, the physical resources of multiple servers, along with associated storage and networks, are pooled in a way that offers greater utilization, availability, performance and flexibility.
Within such configurations, dynamic provisioning allows computing resources to be allocated where needed, and these workloads can even be moved across hardware systems in real time. Scheduled maintenance can also occur without critical applications being taken off line while at the same time reducing overall points of failure.
Virtual machines and virtual infrastructures are being widely deployed, from data centers to industrial automation, video processing and simulation. Trenton offers a broad range of computing platforms that support such applications, using single board computers and backplanes or fully integrated rackmount computers with segmented backplanes supporting multiple SBCs in a single rackmount enclosure.
Two Trenton white papers covering the topics of Cross-Domain Networks and Cluster Computing can provide additional information on how Trenton is serving the needs of high performance computing in a virtualized environment.
This submission is by Mark Lovett, Chief Marketing Officer.
Posted by Tracy Valentine
For the past decade the topic of virtualization has been echoed across a broad range of industries, from video processing to health care, flight simulation and enterprise computing. While it’s often touted as a modern day miracle of technology, the roots of virtualization go back to the 1960’s when IBM developed the first virtualization platforms as a way to logically partition mainframe computer systems in a fashion that would create a number of “virtual” machines.
This was the advent of multi-tasking, in which one computer could run multiple applications at the same time and serve hundreds of users. “Time-sharing” was the buzz word of the day, as the expense of such mainframes required very high utilization rates to justify the cost. In this way, small businesses lacking the resources to afford their own mainframe could still gain access to this computing power by paying for just the time they used.
Personal computers made their way into companies during the 1980’s, and though initially limited in effectiveness due to their stand-alone nature, that situation soon changed with the advent of local area networks. Deployment of these networks provided the necessary framework for development of client- server applications as a way to centralize data yet distribute processing power to the desktop. It was a radical shift in how data was created, stored and utilized, but the architecture was far from perfect.
As it turns out, the client-server paradigm didn’t scale very well, as the cost of deploying separate servers for each application became a costly proposition for most companies. Racks of these dedicated servers were built to handle the workload, but as user counts and the need for processing power and mass data storage increased, IT departments found themselves in constant upgrade mode.
The main problem, however, was really one of utilization. As with automobile traffic on the highway, or electricity usage within a community, there are peaks and valleys in computing demand, which required companies to provide enough processing power to handle peak demand on each of their many servers, even if the majority of the time these servers were underutilized. So how did we eventually get out from under the dilemma of client-server computing?
In part two we’ll examine how the rebirth of virtualization technology on x86 hardware platforms saved the day, and continues to shift the landscape of mission critical computing.
This submission is by Mark Lovett, Chief Marketing Officer.
Posted by Tracy Valentine
The short and “buy our stuff” answer to this question is obviously “Yes”, but the real answer is “It Depends”. Many application factors determine if an industrial computer system solution will be appropriate. A short list of some of these factors include:
- System location requirements
- Projected service life of the system
- Number of option cards needed in the computer
- System expansion requirements
- System power and fail-over requirements
- System O/S and application software solution requirements
The first five items on this short, and by no means all-inclusive list of industrial computer integration factors relate to the computer hardware itself. The last item has an indirect hardware relationship, but for the most part, it is the system’s software component that often times will command the lion’s share of time in integrating a system. If we can unburden the OEM or End User from the task of engineering the hardware and O/S elements into the industrial computer system, that should simplify the integration task while giving the customer’s engineering resources more time for implementing the all important application software solution. For example, industrial automation motion control and distributed process control software solutions can be very complex. Virtualization in a motion control application involves several specialized application software components in addition to the multiple operating systems and Virtual Machine Manager or VMM software. The following articles talk about these virtualization applications and the need to have stable industrial computer hardware platforms for successful installations:
An integrator that understands the subtle nuances of embedded computing hardware can use this knowledge to provide hardware platforms that ensure successful industrial computer installations. The simplicity of the industrial computer’s hardware integration really depends on the integrator’s knowledge of:
- Available processors and which CPUs are long-life embedded processors
- Option card interconnect interfaces such as PCI Express 2.0 and how these interfaces interact with various card types
- The board topology of different single board computers and embedded motherboards, including BIOS configuration issues
- Power supply technology and capabilities such as power availability, start-up surges and redundancy
- Hardware enclosure form factors, air-flow designs, shock & vibration concerns, availability and longevity
Trenton’s product answers to system integration simplification include our line of Standard Systems as well as our Customer-Driven Solutions product line.The TCS4500 is Trenton’s latest standard system and this 4U rackmount computer comes pre-configured with the dual-processor (Jasper Forest) JXT6966 single board computer and a BPC7041 PCI Express 2.0 backplane. The TCS3500 is an example of one of our standard systems pre-configured with a single processor, long-life embedded motherboard. Our Customer-Driven Solutions, such as our 5U TRC5002, enable more hardware choices such as two systems in one enclosure for cluster computer applications. Either way, Trenton simplifies the integration task by taking care of the hardware component, thereby enabling the OEM customer more time to dedicate to the system’s application software solution.
This submission is from Jim Renehan, Director of Marketing.
Posted by Tracy Valentine
Industry terms like cloud computing, cluster computing, cross-domain network communications and virtualization can mean different things in different system applications. While these terms can be useful in describing the basic goals of a system, they can sometimes get in the way of focusing in on the practical hardware and software elements needed for effective system solutions. What's needed to solve many of these system requirements described by fancy industry terms are practical system solutions that are affordable, scalable, flexible and designed to meet both the requirements of today as well as those of tomorrow.
Trenton published a whitepaper that helps you navigate through the maze of virtualization possibilities. The whitepaper focuses specifically on various computer system solutions used in implementing cross-domain networks utilizing cloud computing and cluster computing concepts. On the hardware side of the solution equation, multi-segment backplanes and COTS single board computers provide the scalability, flexibility and network security needed in these systems.
There are several hardware platform solutions described in the whitepaper. These COTS system solutions range from the traditional edge-card SBC to the MicroTCA AMC processor card form factors. The whitepaper describes the combination of COTS hardware and software components that enable multiple applications virtualized at different security levels on each SBC or AMC, and multiple HAP clients that can access each board simultaneously.
Click here for a copy of Trenton's whitepaper on virtualization.
This article was submitted by Jim Renehan, Director of Marketing.