I do not have a car in New York, and I cannot find a reasonable transit route to the data center in Somerset, New Jersey.
This is a small problem compared with the infrastructure work waiting there, but it is the first one I need to solve.
The drive is about an hour. A rental car looks like roughly $50 for the day. I told the client I would normally expense the rental and fuel, but I would not charge for drive time. If they knew a better way to get from Manhattan to Somerset, I was open to it. I am, as I wrote, "a very happy public transit commuter :)."
New York has changed the physical shape of my work. In Denver, a car was assumed. Here I can get around the city on trains, walk to meetings, and keep most of life within the subway map. Data centers do not always cooperate with the subway map.
The consulting job is to get a core data-center environment under control. I plan to visit next week, then make progress remotely while I am back in Denver after the holiday. The work includes virtual-machine infrastructure, resilient network paths, firewalls, switching, and the documentation that turns a pile of connected equipment into something another person can operate.
I am still Director of Technology Operations at Peek. This is additional work, using the same set of skills in a different environment. Peek does not run Microsoft services, but I have years of Active Directory and Exchange experience. I used to manage a multi-site domain trust between an office in Denver and another in Berlin, so a migration between sites is familiar enough to be useful here.
The immediate question is availability.
It is possible to connect the virtual-machine hosts to different switches in an active-passive design. That would not deliver the same throughput as aggregating links, but it would reduce the chance that one switch takes the whole VM environment with it. The right answer depends on what the applications need, what the existing hardware supports, and which failure the client is actually trying to survive.
The request started broadly: help get the data center working properly. The environment produced a practical sequence. Fix the core situation, establish resilient paths, understand the virtual infrastructure, and document the cabling and configuration. Once those are stable, we can discuss the Exchange migration and the next data center.
The sequence matters. It is tempting to start with the visible migration because that is the project people are waiting for. Moving a service onto an unreliable foundation only makes the old delay into a new outage.
The travel sequence matters too. I need access to the room before I can validate what is really there. I need enough time on site to trace cabling, switches, hosts, and power. Then I need a remote path that lets me keep working without renting a car every time a configuration changes.
The servers can be managed remotely after the environment is documented and the access path is reliable. The first visit still requires getting to Somerset, so I am pricing the rental car.
I have been in New York for about six months now. The move was for Peek and the startup market, but it also means rebuilding the practical routines around work. Which email account stays less cluttered. Which meetings can happen by train. Which client sites require a car. How to get back to the city by four because another commitment is waiting.
For this job, I need approval for the transportation expense and a date for the site visit. Once I am in the room, I can trace the network and power paths, review the virtual-machine hosts, document the current setup, and establish remote access for the work that follows.
I will not bill the drive time. I only need the rental car and fuel treated as project expenses.