Calculative Measures to Counter Challenges
A mere decade ago, if when my phone rang at 2 am, it almost always signaled a problem within my data center. Resource intense systems, rapidly expanding storage needs and a rising tide of new technology implementations kept us in a hardware refresh and upgrade cycle that dominated our calendars.
We took the steps most companies took—adopting virtualization, load balancers and storage area networks to stretch times between upgrades and manage manpower. We took it a step further than some and moved our business-critical e-mail system to a new multi-tenant cloud offering. The service that would eventually become Microsoft’s Office 365, freed up considerable resources in our data center and launched a SharePoint implementation, which would replace many of our file servers.
This was a big leap at the time, but the benefits were clear. We followed our goal of economically off-sourcing our data center and hardware maintenance to focus on supporting customer experience. We worked through the loss of control, smaller feature sets and other issues that came with a new service and reaped the benefits.
Despite this change, however, we were again shopping for storage a few years later when we discovered the Storsimple appliance. It offered deduplication and burstable storage that would become our gateway to a hybrid cloud storage area network. At the time, 2011, it worked well with our VMWare systems and offered Microsoft Azure as the only fully integrated cloud storage solution. Determined to find a solution to meet our explosive storage needs, and banking on the reputation of Microsoft to keep our data secure, we took another step into the cloud.
There are often small windows in innovations that a company can use to seize an advantage
Each new technology brings both success and challenges. Storsimple offered tremendous support, which compensated for the then relatively slim offerings of Azure. Our disaster recovery had to be redesigned as Storsimple offered a way to back up data into the cloud but limited recovery options. In addition, users sometimes had to wait a few seconds when accessing large stores of Azure based files.
The success we experienced again overshadowed the challenges. We found it far easier to manage storage and our back-up systems. File recovery that took days now took seconds. Deduplication reduced our storage needs and our ability to seamlessly use Azure file storage gave us virtually unlimited space to meet the needs of our users.
Since then we made only one substantial investment in our on-premise storage environment. We purchased the newer generation of Storsimple, which enables us to recover directly into the Azure cloud. And, like Office 365, the Azure service matured rapidly giving us the support and abundant service capabilities that matched pace with our need.
We now use Azure to host Active Directory and other systems as well. We stand up test systems for short pilots without incurring any long-term commitments. It has reduced our time to implementation as we now buy ready resources, instead of buying over-provisioned systems designed to future proof hardware investments.
With this success, we jumped onboard when Azure’s machine learning offering came online. It was yet another level setting offering in a long line of services that allows small and mid-sized companies the same capabilities once only available to giants. More than just a new service type, it is a new tool to delve into data, understand relationships and perform predictive analytics.
Our results were initially disappointing as we overestimated the amount of data points we could use in an experiment. Using the training routines, and the experience we are gaining, we are making small inroads into the technology, but are still hindered by limited manpower.
There are often small windows in innovations that a company can use to seize an advantage. The window on cloud adoption is still open though it will rapidly become the norm for many enterprises. I believe machine learning is another one of those opportunities. This is a skill, one of many, we must master.
With that need in mind, we are working to shift team members out of back-end support roles and rebuilding our skills with a clear vision toward data and security focused roles.
This level of change is one of the reasons I enjoy working in technology. It is a field, dominated by people that embrace and value change, information, learning, and improvement. It is also an ever-expanding field with a chronic hunger for more workers with new skills.
Add to the burgeoning need an aging work force, potential shifts in immigration, and H1B visa policies, and it will be easy to see why we need to adopt technologies that help transition highly skilled workers into new and growing technical fields.
Large market providers such as Amazon, Microsoft and Google, among others, are already well poised to make this change with us as we seek partners willing to put substantial resources into development and security.
My hopes had always been to reduce our once revered data center to a humble closet. While we are getting very close to that goal, there are still a few on-premise systems we will not move into the cloud yet. Workloads like patching servers with the magic combination of substantial I/O, bandwidth utilization and storage needs are currently still too costly for us to run in the cloud. For the workload we have moved, I sleep well at night knowing someone else is worrying about performance, back-ups, ensuring geo-redundancy and constantly replacing hardware to keep up with growing needs. When I stop to think about it, I cannot remember the last time my phone rang at 2am.