Today, I’m going to be talking about navigating the pros and cons of structured cabling versus a top of tack architecture in the Oklahoma data center industry. I am writing on behalf of the top data centers in Oklahoma
. And for those of you not familiar with who the these companies are or what they do, they are an organization that addresses things like quality, design, environmental, and even safety issues with regard to communication cabling infrastructure. And their primary mission is education.
The main topic of the day is really data centers and, specifically, the data center configuration and why it matters today. Certainly, as data centers become very complex, the design of the cabling system and the topology that’s used become a critical component of how the network performs and how the data center performs. Things like bandwidth, scalability, and virtualization all have an impact on cabling configuration choices. And, conversely, your cabling configuration choices can affect the scalability and things that happen in the data center. Things like power and cooling budgets have to be examined, and cable management shouldn’t be overlooked when we talk about the data center configuration and architectures that are used.
But, what we’re really talking about today is the pros and cons of the different solutions that are available for the data center today. There are several configurations that are commonly referred to when we talk about data center infrastructure design. There is the dedicated network row, where you have central switching for your network. It’s connecting all of the components of your data center and that incorporates a patching area for making those various connections. More commonly, we’ve seen middle of row or end of row architectures, where the network connections are all housed within the middle or end of a row of cabinets and the rest of the cabinets or racks have all of the servers that are running the applications in our data center, or perhaps are part of the storage networks. In those architectures, there is a structured cabling system that’s used to connect the various cabinets or racks that house the switches as well as the servers themselves. And, generally, the network components are housed in their own cabinet or their own rack either in the middle of the end of the row of those servers. Those are the more common configurations that we’ve seen for probably the past five to six years.
One of the more recent introductions in the architecture was the introduction of the Top of Rack configuration in which the switches are now located at the top of each rack or cabinet of servers, and there’s a direct connection between the switches and the servers, and then all of those switches are connected back to your core network via optical fiber.
But, the point I want you to walk away with today is that there really is no single end all solution. None of these architectures is always right for every data center. And we’re going to look at the different configurations and their pros and cons.
So, I mentioned several times in that middle of row or end of row architecture the term “structured cabling.” And, in the data center context, structured cabling really is designed as patch panels that are connected with permanent or fixed links to the switches and to the servers in the different parts of the architecture. But, you’re making these connections between the switch and the server via a set of patch panels that represent the server ports and the switch ports. Typically, those patch panels are located either at the end or the middle of the row, as I mentioned, and they’re in a separate cabinet or rack. All the changes that we’re making between the different application servers and the network sub-segment that they need to be connected to are made via this central patching area using patch cord, maybe a fiber optic patch cord or a copper patch cord.
The real goal of this configuration or this structured cabling is creating an any to all configuration, which really means that you’re able to route or connect any application server to any switch within the network, specifically, switches that may be part of a sub-network that needs to be managed separately.
The TIA-942 Data Center Standard really promotes the idea of utilizing structured cabling in various sub-segments of your data center design. It really defines the minimum requirements that are needed to support all of your applications and all of your network types. It actually has a built-in redundancy mechanism that would be supportive of growth and defines specific functional areas in which you have the patching and the distribution of the different connections located in one area and all of your sub-segments in the — or all of your network equipment connected separately. But, really, the goal of the TIA-942 Data Center Standard was to create a common architecture that would be flexible enough to support all of the different types of networks, all the different types of applications, and some of the growth that we’re going to be anticipating because data centers are very much a dynamic environment. The data centers and the 942 Standard really are focused around a structured cabling approach. And this is they way it is recommended by the experts today.