This audio was created using Microsoft Azure Speech Services
At this point I’m sure most of us have heard of edge data centers. And in the case of businesses that have many distributed locations (e.g. retail, manufacturing, industrial), it’s not financially feasible (nor recommended) to upgrade all their sites at once. In a previous presentation (and white paper 256), Kevin Brown, Schneider’s CTO, described a logical way of identifying which edge data centers are most critical, and therefore, which are most in need of an upgrade.
The next question then becomes, how do I go about improving the availability of my edge computing solutions? That’s what prompted the writing of a new white paper 280, Practical Guide to Ensuring Availability at Edge Computing Sites. This white paper is written in simple terms with many practical and actionable availability improvements to your data center design for edge. In fact, it’s broken up into a series of checklists that you can print and use on a walk-through of your edge site.
Download White Paper 280:
Practical Guide to Ensuring Availability at Edge Computing Sites
Practically Assess Availability of your Data Center Design
The checklists are broken down into eight systems:
- Power
- Cooling
- Rack
- Physical security
- Fire protection
- Environment
- Network connectivity
- Management
The improvements are set up qualitatively such that the improvements at the top of the list give you the biggest availability improvement for your money. The improvements are ranked relative to a “baseline” which I define as a stack of IT equipment, sitting on the floor of an open room, powered directly from utility, with only comfort cooling (i.e. very low availability). So, from a power perspective, what could you do to improve the availability of this IT stack? You get your biggest improvement by plugging the IT equipment into a single-phase UPS. Usually a line-interactive or standby UPS. For the money, you get a huge pickup in availability with this one improvement.
Now what’s the highest priority cooling improvement? It’s pretty clear that if the building’s comfort cooling system turns off on the weekends and at night, you need a way of keeping the IT temperature within ASHRAE guidelines. There are several ways of doing this and not all require you to spend big money on a dedicated compressor-based air conditioning system. (Our white paper 68 helps you determine this just by knowing some basic information about your site!) Throughout the checklists, I also try to explain how a particular improvement addresses availability – which most are ways of reducing the likelihood of human error.
Assessments That’s Right for your Edge Data Center
On the spectrum of complexity, I’d like to think that this checklist method is the simplest way of improving your edge data center availability. However, as I state in the paper, if your business risks losing a significant amount of money due to downtime of a particular edge data center, it may make sense to invest in a quantitative assessment. These are typically called probabilistic risk assessments (PRA) which can cost upwards of $10,000 for a small edge site.
To access the practical checklist you can use to evaluate the availability of your edge data center, download white paper 280, Practical Guide to Ensuring Availability at Edge Computing Sites.
Leave a comment below or check out other blog posts from the Data Center Science Center team.
