PS: Sounds great. Lets look at Datacenter Optimization. In the datacenters today we find inflexible infrastructures. Physical servers that you install the OS and applications on, and making it difficult to easily move between platforms, or between physical and virtual machines. They are bound to the hardware that they were installed on. As a result, we find that the enterprise datacenters are running in a sub-optimal manner. Servers with many resources are handling small workloads; they’re underutilized. Some servers with small resources are handling very large workloads; they’re overutilized.

If you could measure the workload and resources and then move or migrate the workload easily within the datacenter, whether it’s between physical servers or virtual infrastructure, you could more easily optimize by matching resource supply to workload demand. The underutilized server could be consolidated; the overutilized server could be deconsolidated, resulting in a more balanced datacenter.

So what PlateSpin has done is create a technology we call OS Portability, which allows you to decouple the operating system, applications and data, and move them around from physical servers, virtual machines and image archives.

How does PowerRecon 2.0 fit in? It gives visibility to all the resources in the datacenter. Many datacenters don’t know how many servers they have, they don’t know what’s running on those servers, they don’t know what kind of workload utilization those servers are running off, or how well the workload demand and its resource supply fits it. So, what PowerRecon does is allow the datacenter manager, or the datacenter user, more visibility into what they have and how effectively workloads are assigned to resources.

Some of the main benefits and features of PowerRecon 2.0 are very rich data modeling of organizational resources and workloads that gives the user insight into how things are running. Customized reports identify servers that need to be consolidated or deconsolidated. The product is installed in the datacenter and is agentless, which helps you avoid security violations since there is no need to send the information or data to a third party for analysis. We also have this notion of asset inventory. All the hardware and software components within the datacenter are detected along with their workload and utilization specific details. So, from a server consolidation point of view, being able to balance the workload demand and resource supply, getting visibility into the datacenter, and doing the capacity planning, PowerRecon 2.0 really fits in to make that a lot easier.

Here is a basic workflow of how it works. You install PowerRecon 2.0 Server and enter your admin credentials. PowerRecon first collects all static information, such as what kind of machine it is, what make, what model, what brand, what kind of hardware it’s running on, such as the CPU type, number of CPUs, memory, disk and network resources, and so forth. Then it goes along to collect performance data. Depending on what the user wants and how long they want to measure their servers, it will automatically go out and measure the utilization for all the servers that the user selected.

The data is dynamically collected, sampled and stored into a local database on the PowerRecon 2.0 Server. It is then summarized and reports are generated.

PS: That’s right. Power Recon goes off and gets the static inventory data of the particular servers you select to monitor. It will go out and measure those servers for whatever amount of time you want, at whatever interval you want, and then it will generate visual and graphical reports. It will also generate sortable reports so that you can figure out the server utilization in terms of CPU or memory, for instance.

PS: We’ve designed this product more along the lines of consolidation and optimization facility. Rather than compete with the MOM’s and the Tivoli’s of the world, where it gives you red light / green light based on utilization, send off alerts, and SNP traps, we’ve designed this mainly for analysis and assessments, for the time being. We’re going to be looking at enhancing this for the disaster recovery space. For example, being able to analyze a server and its workload. If its workflow goes up to 99.9% - do something. Maybe cause a failover to occur and so forth. But right now, it’s designed as an assessment and analysis tool for server consolidation.

PS: In addition to the numerical reports that you get, which would be an average or the standard deviation analysis, you can also do a time-based analysis. Basically, it gives you a time-based characteristic over a 24-hour period.

Another type of graph it can produce is a stacked graph, which essentially, cumulatively gives you an area graph of what a server would look like after it’s been virtualized. It can also create different graphs such as curve graphs, candlestick charts and so forth. The stacked bar chart is useful because it indicates the blank space between workloads and helps you to stagger the workload on a time basis.

We can also look at disk utilization and memory. The memory stack shows you how much memory is needed on your virtual machine. Network utilization, stack package per second and disk utilization all give you a good insight into how to stack servers into a virtual machine environment and how to consolidate from that perspective.

In the next 45 days we have planned to offer a Consolidation Add-On Module, which allows the user to press a button, and based on their criteria, automatically suggest and identify the servers that are to be consolidated and where they should be moved to based on workload and resource characteristics.

PS: It will only be available for PowerRecon 2.0 as a downloadable add-on module and uses your existing data. You press a button and it automatically spits out suggestions on where to move things. Like you said, it takes a lot of work out of the analysis phase. From there you can use PowerConvert, drag and drop servers into their blade or virtual environment.

PS: Well, we have a performance option piece that allows you to adjust the interval and frequency at which monitoring samples tape in order to minimize the network impact based on the number of servers you have. Obviously, if you have more servers, you want your interval to be larger between samples. If you’ve got a smaller amount of servers, and you want to get a little more granular with the detail, then you make that interval smaller.

PS: What I went through with PowerRecon is really the top two pieces of the equation; awareness of resources and workloads, and how they fit together. The other product that we have is called PlateSpin PowerConvert. This is what we deem the OS Portability technology, where we take an operating system, all the data and applications, and move it from anywhere to anywhere. We decouple it and we can move it from physical machine to physical machine, physical machine to virtual, virtual back to physical; there really is no limit as to where you can move it. Basically, you’ve got this notion of being able to lift a server off; get the entire software stack and move it from anywhere to anywhere. In one of our recent studies, our customer AXA was able to virtualize 300 physical servers into VMware ESX servers, from Tokyo, Japan to Melbourne, Australia. And all of this was done from New York City.

PS: I like to use the analogy that PlateSpin PowerConvert is like the Star Trek transporter beam. Where you energize a server from place to place. Only we don’t move the hardware, we just transport the software stack.

PS: No, you can remotely access the machines on the network, enumerate them, discover the inventory information, and then choose which ones to measure in terms of performance.

PS: It can do Windows on a physical or virtual platform; it can do Red Hat LINUX as well.

PS: Only Red Hat for now; we are working on other flavors of Linux.

PS: PowerConvert has a simple user interface. Like PowerRecon, we enumerate the servers on the network agentlessly, and discover the server detail. In addition to the details you would get from PowerRecon you also see the services that are running as well. If I wanted to virtualize Server X, or move it into a blade server, or capture it as an image, I would simply move it across to my ESX Server, drag and drop –again, this can be done completely remotely, and then what happens is I get a target job configuration type of dialog box. What this does is allow me to keep a source machine offline after the conversion. I can schedule the job, enter my admin credentials, and get notifications on the job progress. I can also do things like allocate a certain amount of memory from an ESX perspective, CPU allocation and so forth. If there is a multi-processor machine on the other end and the ESX Server is running on a multi-processor system, I can assign the CPU’s of that virtual machine to a particular physical processor.

PS: Currently, we need to take control of the production server to copy the data; it would need to come offline.

Live transfer is planned in the next release of the product, which allow us to get into the disaster recovery marketplace. We will be able to take a snapshot replica of a physical machine, move it into a virtual VMware ESX Server or to a blade server, or another physical server in which it would act as a warm standby in case anything happens to the primary. We are able to configure the network to turn off. I can make them a manual or automatic start. We can adjust hard drives to select specific drives or files.

Although I’m not actually connected to a live system now, essentially you’ll get the idea that you can save it as a job or you can just press start. After start, you go to the jobs view that lets you see the progress in which the conversion is taking place. And this steps you through everything that it’s doing; creating a VM, copying all of the images over, capturing the image from the source machine and so forth. When it gets to step twelve, you are pretty much done. So essentially, the workflow process is to connect to the network, discover your server, drag and drop, configure the target machine and press start. Go away for dinner, come back and you’ve performed your conversion.

Looking at our product offering at this point, we are looking to complete our equation. We’ve got this thing that identifies servers and where to move them. Then we’ve got this other piece that actually moves them and automates that process. How people would use PowerRecon and PowerConvert in the data center? With PowerRecon you press a button and it comes up with servers to consolidate or deconsolidate, depending on the workload and resource matching, and PowerConvert goes off and does the migration.

We envision datacenters using Power Recon every three to six months to make sure that they are continuously optimized. What we are planning is put together a complete end-to-end Server Optimization Suite solution. Eventually our model is to have it completely automated, so that you just press a button every three months or every month to make sure that things are optimal. Eventually, they won’t have to press the button anymore. And servers will be moved from the datacenter to where they will best fit in terms of workload and resources. What you end up with is a closed loop solution where I can be aware of how workloads are running.

What you would see in a datacenter is being able to move servers across different infrastructures in an automated manner with an end goal of having perfectly matched workload and resources across the datacenter.

PS: Physical or virtual. It actually knows what a virtual machine is and what the hierarchies are on the host systems. So, its very much Virtual ware.

PS: That’s right. It will be able to identify all the servers and make suggestions as to where you should move them. Then that would be shuttled into PowerConvert and again, you would just press start and all those conversions can take place.

PS: That’s right. After datacenters have done their initial consolidations, they will find that servers are going to change their workload. Over time, their workload is not quite as balanced as they could be anymore. And so that’s where this product suite will help them on an ongoing basis as well, in addition to the one-time server consolidation initiative.