Data center planning: Surprising tools of the trade

The last time you applied for a tech job (or evaluated a résumé of someone applying for one), how much weight did you apply to expressing your skill with basic productivity tools like Microsoft Office or OpenOffice? My guess would be very little. In the résumés I’ve seen from prospective applicants, it’s sometimes not even mentioned. After all, being conversant with tools that are such basic parts of life in business today might seem to be a given.

It’s anything but. I’ve met and worked with highly skilled network, server, and storage engineers who have only a passing knowledge of Excel and Visio (or their respective OpenOffice equivalents). Although it may seem a little silly at first, I believe these skills are almost as critical as having a solid understanding of the infrastructure tech you’re tasked with maintaining.

I spend an enormous amount of time working with Excel and Visio. Rather than just creating documentation for a job that’s already been done, I routinely use them in the planning process, communicating design changes to business stakeholders and, in some cases, as a means for configuring hardware and services. I can honestly and unabashedly say that without Excel and Visio, I’d be almost unable to do my job.

Spreadsheets to the rescue

Although they’re often considered the domain of finance folks, Microsoft’s Excel and OpenOffice’s Calc can be massively powerful tools that can fulfill a wide range of roles. Given the task of planning a new data center hardware deployment, I use Excel to plot hardware budgeting, engineering time/task allocation, IP address assignments, power and cooling calculations, cable length determination, number of needed fiber modules, SAN LUN sizing, full, easily modifiable switch configurations, and more.

In fact, I spend the first 20 to 30 percent of any project bashing information into Excel. By the time I put my hands on a piece of sheet metal or a cable, I have a ready reference for how each piece should be labeled and where it should go, and I’m reasonably certain I’m not missing anything. Compared to the ad hoc way I used to do things, this saves me an incredible amount of time — both in terms of trying to remember how things should go together and in reworking them when I inevitably realize I got something wrong the first time around.

Let’s say I’m helping a client migrate its VMware-based virtualization environment to a new primary storage infrastructure. The first thing I do is fire up the vSphere client; navigate to the list of virtual machines; add extra columns to the view to include IP addresses, DNS names, and the like; and export the list to a CSV file (a capability that many tools have). Once I have that list opened in Excel, I can total the amount of storage consumed by and provisioned to each VM, perform quick calculations to split them into different LUNs, and know exactly how large each data store I provision on the new SAN should be and how full each will be after the migration — all before I even get it out of the box.

If the job requires it, I can add a few columns to include replication, snapshot, and tiering SLAs and have the client go through the sheet to make sure it reflects each individual machine’s criticality — further guaranteeing that every machine will get moved to the right place the first time.

In another example, I might be configuring a pair of top-of-rack switches to aggregate a new virtualization and IP storage infrastructure. In that scenario, I need to know how many cables are required, how long they should be, which ports each should be plugged into, and how those ports must be configured. Again, I start with building a sheet in Excel that shows each switch port, its configuration (VLAN, flow control, MTU settings, and so on), which device port it’s attached to, and even which rack unit that device is planned to be in.

With all that information in place, creative use of Excel calculations allows me to know exactly how many cables I need, exactly how long each should be, and even the color of each extensions (when the client wants cables color-coded by task, that is). I can also have a column that uses the port configuration breakdown to build out a full layout for each interface — so deploying the switch is a simple issue of copying and pasting the prebuilt configuration, complete with accurate port descriptions, directly into the switch’s command-line interface.

Although you can certainly complete both tasks (and many others) without that level of prework, I guarantee you’ll make mistakes and/or end up with a configuration that you wished you had done differently after seeing it all go together. The spreadsheet planning exercise avoids those unfortunate results.

A picture is worth a thousand words

When it comes to designing networks, planning the construction of a new data center, or just determining where new equipment will go in an existing rack infrastructure, Microsoft’s Visio or OpenOffice’s Draw are my go-to tools. Among engineers, these tools are known for creating network documentation, but they also can be excellent planning aids — allowing you to quickly communicate what an implementation is going to look like to stakeholders and to identify potential problems before you find yourself staring them in the face.

Rack diagrams in particular are extremely helpful, and not just because they make pretty pictures or might help you direct a crew that’s getting the rack and stack done. When planning fast-growing storage and server infrastructures, it’s helpful to know how many disks or blades you can jam into a given rack before it gets full — and how best to plan the placement of equipment so that you won’t have to move it again later. In larger projects, the same can be said for the floor plan of an entirely new data center — complete with the locations of needed circuits and HVAC zones.

The end of post-project documentation regurgitation

Although office-productivity tools can be enormously helpful in the planning stages of a data center project, it also handily absolves me of one of my least favorite tasks: creating project documentation. There’s nothing I hate more than spending time at the end of a project going back and regurgitating onto paper what I and my team have done. It’s not only tedious, but documentation created after the fact is likely to miss important information and unlikely to be a useful resource in the future as things change or new systems are added.

However, by having everything planned to a T before work even begins, the project documentation is largely done prior to the start of the project. Plus, you have a prebuilt planning tool for the next time you add more systems.

If you find yourself among the ranks of engineers who don’t regularly use these old-hat tools to their fullest, give it a shot. You may find, as I have, that they make life far easier, provide you with documentation that will be useful years later, and serve as excellent communication tools when you’re discussing your infrastructure with management.

This article, “Data center planning: Surprising tools of the trade,” originally appeared at InfoWorld.com. Read more of Matt Prigge’s Information Overload blog and follow the latest developments in storage at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.