
Welcome to Darren Bonawitz’s weekly podcast.

By: Darren Bonawitz
Hi. This is Darren Bonawitz again, with 1102 GRAND. I’m one of the owners of the facility. For those of you not familiar with 1102 GRAND, we are one of the Midwest’s major Internet hubs and co-location facilities. This series of podcasts talks about different industry trends and topics that I find interesting as I stay up on the news and discussion with people in the industry.
Today I want to talk about what I call “low-power consumption and high heat servers.” These are the servers that are coming out that use less power and cooling in a data center. Generally, the ones I’ve been seeing are based on AMD’s Opteron and Intel’s Atom Processor lines. The reason I want to talk about them today is because I think they are a solid first step, as data centers are looking for ways to cut their carbon-footprint, or become more energy efficient. In the past, the metric that really mattered in the “server world,” was Total Processing Power. No matter what, that’s seemed to matter. Every new series of servers that came out really had to deal with how much processing-power it had, but if you look, especially as things with personalization evolved, one of the things that makes virtualization so interesting is that CPU usage on the server is generally low. So, what I think is interesting with these processors is that they are still powerful processors that use lower power, but at the same time, they’re going at it from a different angle. Really, what they are trying to do is figure out a way to increase performance per watt of power used. So, I think it’s really something that’s really relevant in today’s world, with “green” being such a big topic in datacenters. The thing I wanted to make people know is that while these servers are a great first step, just keep in mind that there is not going to be a single “silver bullet” that’s going to solve the issue of controlling heat in the datacenter. With that said, processor temperatures are definitely one of the big targets, and they are certainly one of the major culprits of the heat-generation. This is a solid first step. There are also new technologies that may be more “liable.” I guess that is what I would call it. That has to do with technologies that I would call “fail-in-place servers.” I’m not going to go into that today, but if you have time, go on and Google it. Check “fail-in-place server platform,” and read-up, and that’s something that I’ll be doing in the meantime as well, and we’ll be talking about how to do this. See when it makes sense and when it doesn’t make sense to use these servers.
The one thing that comes to mind for me today are legacy-datacenters that don’t have the ability to handle high-density heat loads. These servers, because of their cost and their price-point, make it so you can actually divide these servers and span them across the series of cabinets, and spread your heat-load and prevent heat-spots. I think that’s definitely something that folks that are battling cooling issues, but still have some floor space, can use as a possible avenue to combat this. If you have other ideas, or if you’ve been using these servers and could give some feedback, please let us know. As always, we welcome your feedback and future ideas for topics to discuss at podcast@1102Grand.com Thanks a lot, and have a good day.
Tags: 1102 GRAND, Bandwidth, Bandwith, call center, carrier hotel, carrier neutral data facility, Carriers, choosing a data center, cloud computing, Co-location, co-location services, collocation, collocation cross connect, Colo, Darren Bonawitz, Darren Bonawitz podcast, data center, Data Center Security, data circuit, data management, Data Suites, efficiency, electricity, environmentally conscious, Green IT, green tech, green technology, greening it, hardware, hardware maintenance, heating and cooling, Internet Hub, Internet hub and data center, IT, IT infrastructure, Kansas City data center, LEED, Midwest data center, Midwest Internet hub, raised floor, remote facility, temperature control, VoIP