Strategy | The Data Centerless Business
Case Study – How to move a traditional, legacy bound operation to an “internal” cloud service provider and reduce IT cost to below 1% GP.
In today’s operating environment, having the infrastructure down the hall in an on premise computing center is no longer a necessity. Gone are the days of going to the printer room to retrieve a green bar printout run from a legacy program.
Now, with the help of an effective IT consulting firm, remote attachment to a server in France is almost as responsive as on a local system in the U.S.
Cost, size, heat, speed, and reliability have all played a role in creating state-of-the art computing solutions that rival rooms full of equipment. The reduced complexity, ease of service and support, built in utilities and capability automate many of the functions that required an entire floor of IT staff, and can drive down cost.
With increased reliability, improved SLAs and support, we can finally turn loose of the traditional computer room mindset. Dramatic improvements in networking and WAN speeds have made it possible for servers to be across town, the next state, or even another country.
With this broad spectrum of capability the quest for lower operating cost and higher reliability have become a reality. IT departments today, on average, operate on budgets less than 1.5% of Gross Revenues; however, with the process deployed cost was reduced to below 1% of GP.
This favorable operating model did not jeopardize customer satisfaction, and maintained an uptime percentage of 99.5%, all this, and without a helpdesk.
By automating our helpdesk tool, in this instance a product named HDPlus was used where escalation rules were created and put in place to manage every ticket entered. Any ticket created that remained in an “unassigned state” for over 2-hours automatically escalated to the team lead, then began a journey up the management ranks if it remained unassigned, potentially to the CIO. Every server and even user desktops were monitored. Rules established that monitored CPU utilization, memory, disk capacity, etc. triggered alerts. The alerts in the form of emails routed to the helpdesk system where they were parsed for key words. The “key word” driven process rules resulted in tickets automatically created and categorized as: high, medium, and low priority. The parsing also identified equipment type and problem description, these variables triggered a rules engine within HDPlus that determined which team / group to assign the ticket.
The above mechanism kept ticket management and SLAs at the stated level.
Coupled with this strategy, IT also provided training programs with the user community that were routinely held once a month on subjects such as: how to self-diagnoses and self-service various issues. Our subject matter experts used ticket types and the volume of tickets from the previous month to drive the agenda. The overall impact drove ticket management constantly on a declining trend. The ticket system also permitted the ability to show similar tickets of potential resolutions as users entered their own ticket, many times this resulted in a user canceling the ticket they were creating and solving their own issue.
With service and support squarely resolved the focus shifted to reducing cost in physical data center operations. Data centers were outsourced to local vendors and equipment was never purchased, but leased.
With this strategy, maintenance cost remained low and because every system was within the 3-year warranty period and was less than 3-years old, failures were reduced to near zero.
The network, IP phones, firewall management, and WAN service and support were also outsourced. Routers and firewalls also monitored by the internal monitoring solution, and built-in escalation mechanisms caught virtually every problem over a 3-year period of operation. There were no unplanned outages.
Server counts were reduced by 80% via a very active and aggressive vitalization strategy. Tier two (not enterprise) storage was utilized from major vendors, and call-home strategies deployed. Service on storage related issues were proactive, and handled during evening hours minimizing the impact on production schedules, if at all.
Disaster Recovery was accomplished by mirroring the two major sites. One in Indianapolis, IN. and one in Augsburg, Gr. Each site utilized a Data Domain device to automated continuous backup operations. The backup copies were mirrored between sites on 4-hour schedules, thus placing a complete “re-startable” backup copy at the remote site, and a side benefit …No Tape Backups!
Twice a year SAP was brought online in Indianapolis and components manufactured at the Germany sites for a 12-hour period, then the systems were brought down and returned to normal configuration. On the alternating 6-month schedule a similar process was executed, but in this case Germany did the hosting.
Careful planning, responsive system design, and innovative thinking helped achieve a consistent less than 1% of GP cost model. Our big data consulting firm can help you significantly reduce your IT operating cost too. Contact HDPlus today to learn more!