Friday, October 03, 2014

Work Arounds




Over the last few months we have been rebuilding our network infrastructure which has been evolving with our business over the last 25 years. The last major upgrade we did was in 2002 and since then there have been a lot of additional servers and people added to the company. As we grew and added to our network we kept adapting what we had to make it work, so there were many work arounds and slowly but surely we grew a very complicated tapestry of a network design.  Last year we decided that our engineering team would slowly rebuild our network infrastructure and storage design in between working on customer issues and storage implementations. In addition, we would change our phone system (VOIP) also,  but we manage the phone system as part of Administration and not part of our Network. 

Once we determined that a rebuild was required we had to decide what was critical infrastructure and what order things could be worked on. Based on what we learned about the new insurance rules on Cyber Liability and Data Breach we determined that we had to keep all data and email functions in house to remain in compliance with the insurance rules and the contracts we had with customers in regards to data security. 

Changing our Physical Servers into Virtual Machines was one of the first things we did and the project went well. We were able to reduce our footprint in our server rooms and we learned some tricks of the trade that already have helped us in supporting our customers who are working in Virtual Machine environments. In the process of virtualizing we were able to clean up quite a few accumulated workarounds in our server architecture and we modernized quite a few of the bits and pieces of our network.

Next came the task of upgrading our routing and switching between our locations to help with our Disaster Recovery and Business Continuity plans.  As our network had grown over the last 12 years we had added some interesting work arounds that allowed applications to talk to each other and created subnets for projects that no longer existed. Throughout the process of re engineering our network we had to keep considering what we had to be prepared for over the next few years in business and as we got closer to completing the project a customer had a request that we had not considered and we needed to consider whether to start the work around process again or purchase new equipment to handle the new customer request. 

We also decided to outsource our phone system and got rid of our Nortel Meridian system. We researched a lot of providers and when we decided on the provider to go with we still did a lot of testing and tried to anticipate everything that our company and our customers would request. Even though we thought we had covered everything we would need, within a week we had to figure out a way to insert a code into the displayed number to indicate the call routing to the individual user along with a redesign of the way our follow me and hunt groups worked.

Despite our attempt to remove all workarounds, it seems we are constantly having to adapt one part of our Network, Servers  or Storage to some new trigger or requirement, whether it be a virus, malware or a new customer request or service enhancement that we are going to provide.  As a company we are all pretty adaptable to new situations and customer requests, and as we rebuild our network it has become clear that no matter how much we try to build a logical network design, inevitably we are going to have to have some workarounds in the tapestry that is our Zerowait infrastructure.


No comments: