GUEST BLOG BY EDGAR HAREN...When you look at the layout and design of today’s data centers, you’ll notice a drastic contrast between the data center’s of large corporations and those of SMBs. In fact, there are even major differences between the large internet, software and service providers and other larger enterprises. On one side, you have large organizations where the bulk of their IT infrastructure is their business. On the other, you have firms that see IT more as a cost center. These companies — such as Facebook and Google — have state-of-the-art data centers, often placed in strategic locations where they can utilize specific resources to help reduce their power and cooling costs. These firms also tend to design and implement their own IT infrastructure. A step closer to the middle are large capital firms who have more standard data centers with branded compute and storage IT elements whose primary focus is on reducing their capital expenditures (CAPEX) and up-front acquisition costs. Finally, there are SMBs with a polar opposite “data center” experience where they are often forced to convert traditional office space into a makeshift IT footprint.
For smaller firms, investing in servers, storage, networking and software can be a painful experience as the primary financial stakeholders may not understand the necessity or importance of this expensive investment for their daily operations. Many smaller companies also rely heavily on value added resellers (VARs) to recommend IT solutions. As a result, the key decision makers may be further removed from the IT requirements that are necessary to keep the company productive and meeting customer demands. For this reason, many smaller firms keep IT investment to a minimum and avoid additional costs such as server infrastructure and options. The reality is that these solutions — such as server racks, power distribution units (PDUs), uninterruptible power systems (UPSs) and other options — help organize, protect and sustain daily operations. The irony is that although these infrastructure solutions often encompass less than six percent of a customer’s overall IT costs, many end users don’t see the value of such an investment until their systems are compromised. In a study conducted by the Ponemon Institute, 24 percent of primary root causes of unplanned system outages were caused by human error. The primary impacts of these outages were business disruption, lost revenue and reduced end-user productivity — where as equipment costs were the second to last variable.
I recently discussed the topic of contingency planning with my friends at Eaton, a global manufacturer of server racks, UPSs, PDUs and other server infrastructure solutions. We came up with a short and simple list of items to consider.
Server Racks – Make the Rack Part of Your IT Stack
- Secure your valuable IT hardware.
- Elevate servers and other IT assets off the floor away from potential spills and minor flooding.
- Provide a means to optimize the organization of network and power cables to ensure proper identification and improved front-to-back server cooling. This will also reduce the potential for accidental unplugging.
- Allow for the most efficient use of precious real estate and floor space, while protecting against seismic events where required.
UPSs – Backup Isn’t Just a Storage Term
- Server-grade UPSs provide line conditioning to ensure your attached IT hardware isn’t impacted by brownouts, blackouts or unclean line signaling. This ensures that the hardware’s internal components are operating at full efficiency and are not compromised by spikes or sags in power.
- Configure UPSs to ensure that your staff or outside consultants not only have enough time for a proper shutdown and data backup, but to also provide a means to save critical data and work in progress.
- Ensure that both physical and virtual servers can be prioritized for graceful shutdown, allowing administrators to place a hierarchy on systems and subsystems running the most-critical business applications. (Eaton’s power management software allows you to do this.)
- Provide meaningful power reports for facility and IT administrators to evaluate power consumption and power integrity to ensure all IT systems fall within the power envelope of the facility, and that power anomalies can be addressed to prevent damage to invested IT hardware.
PDUs – There Is Nothing Wrong With PDA For Your PDU
- Rackmount PDUs provide expandability for power connections to allow for the most efficient use of floor space and power infrastructure. PDUs also assist in creating a redundant power layout to ensure that mission-critical systems are not fully dependent on one AC feed.
- Intelligent PDUs can provide remote boot/reboot to locked systems, while also providing key power metrics per outlet. This is ideal for monitoring power consumption, and for chargeback situations in cloud environments.
- PDU software can assist with asset management to reduce the propensity for human error.
- To further assist with downtime caused by human error, Eaton PDUs also feature a secure AC power cable retention mechanism, designed to prevent the accidental unplugging of AC power.
In order to help companies determine their risk exposure, Eaton provides a very helpful calculator to assist decision makers in quantifying the economic impacts of data center or system downtime. My friend and I put together a quick reference solution for remote offices, retail outfits and small-scale IT departments. This reference architecture would be ideal for running retail software, creating a virtualized multi-tenant environment, (between 48-96 VMs for two servers depending on memory required), or for traditional IT applications such as file, print and backup activities.
A) Eaton basic PDU
B) Network switch
C) 2 x Lenovo ThinkServer RD640 rack systems
D) Lenovo PX12-450R storage array
E) Eaton 5PX UPS
F) Eaton 5PX extended battery module
G) Eaton 12U server rack
Hopefully you found this blog informative. I’ve included some links below that you may find useful. If you have any further questions on contingency planning and server infrastructure solutions, please contact me, Edgar Haren, at firstname.lastname@example.org.
- Eaton Cost Of Downtime Calculator
- Eaton Rack & Power Landing Site
- Ponemon Institutes Research Entitled “2013 Cost of Data Center Outages”
- Eaton Blackout Tracker & 2013 Blackout Tracker Annual Report - According to Dunn & Bradstreet, 59 percent of Fortune 500 companies experience a minimum of 1.6 hours of downtime per week. Assuming an average staff of 10,000 employees who are paid an average of $56 per hour (including benefits), the downtime loss in labor alone for a Fortune 500 firm would ring up at $896,000 per week — or more than $46 million annually.