For many, many years, I have been a pessimist in supporting the migration of my applications to the cloud. I had spent the last 25 years of my career running data centers of my own where my staff and I would architect, design, buy, build and run hundreds of systems to support the company’s business. At one point in the late 90’s, I was running more than 8000 servers across multiple data centers with an annual run-rate in the millions of dollars.
Why was I a pessimist you ask? Well, it was always the standard cliché that you read about in any study. I felt I could do a better job than one of those “big” companies at managing my own environments and the amount of redundancy, availability and monitoring I needed wasn’t available at the time. I had also felt that I had perfected the “speed to market” issues by having a small quota of spare equipment ready for a quick start-up if needed along with standard images setup to boot.
About four years ago, I started consulting for companies. I was hit head on with the cloud issue as many of my customers wanted to migrate to the cloud as they needed to reduce their costs and overhead of having data center floor space, assets on the books and large IT staff and tools to manage it. I was forced to begin trying it out, having to take crash courses on the differences between IaaS, PaaS, SaaS, private cloud, public cloud, etc. Oh, I had implemented virtualization with VMware and other products to maximize the efficiency in the hardware at my data centers, but still had not released on using the real “cloud”. I spent the next three years working with a hundred or so clients coming up with varying types of solutions for them in the cloud from all different industry sectors and helping them with their migrations.
I then moved on to Experts Exchange where I am currently at. When I arrived, they were in the same situation I was in early on in my career and as all of those consulting clients were. They had a large implementation of servers of varying configuration and age with a need to scale for new projects and functionality, customer growth and to decrease their overall costs of IT. There were some critical dates coming up very soon that included renewals of colocation facilities, software licenses, hardware maintenance renewals and the need to buy new equipment.
We quickly started to investigate the ability to move our systems to the cloud. I dedicated two people to prototype our application in the cloud. Some of the staff had already been tinkering with Amazon Web Services for their personal use, so I gave them the challenge to test it within a week. They were able to prove it out in three days. We made the decision to go and were fully migrated within two and a half months.
The migration was actually the easy part because we had some good planning done by the IT staff. The most challenging part was to determine what tools we were going to use to manage the new environment. We had been so used to our customized tools and/or off-the-shelf open-source tools, but the best thing about the migration was that there is a huge menu of options to choose from to manage environments in the cloud.
It all depends upon how much you want to spend and how much you want to manage. There are always third party companies that you can pay to manage your environment in the cloud, but we still wanted to have a little ownership in this. We found tools like OpsGenie, DataDog and the standard AWS console as the most useful tools to manage our environment; in fact, they do a better job than our previous tools did before and allow us to even have some customization capabilities. And, we send critical events directly to our pagers and non-critical to our email; all for a “pay as you go” monthly fee.
What made the migration to the cloud a success? Having the right IT staff that were willing to take a chance and were given the latitude to make it successful. Should I have migrated to the cloud? Absolutely. The cloud today is ready to support just about any workload that could be thrown at it. The cloud has provided us a much more scalable and responsive infrastructure than we ever had doing it ourselves. The tools are awesome and the availability of our systems has been well north of three-nines for the last four months. My IT staff used to get as many as 20-30 alerts per day for problems in the environment from 5pm to 8am the next morning. We are now lucky to get as many as 10 alerts in an entire month and then those are mostly FYI’s.
And, the most important aspect of our success is that Experts Exchange is now positioned ourselves to deliver features and functions to our community members at a much faster pace. For example, each of our community members now have what is called a Content Feed on their Home screen when they login. That would not have been possible so quickly without moving to the Cloud as we were able to quickly take advantage of some critical caching technologies available to us that we would have had to build on our own had we not migrated. And, we are working on improvements to that Content Feed that will be due out very soon.....all due to our move to the Cloud.