How to Get Internal Buy-In for an IT Infrastructure Overhaul
Wibmo is a leading technology and service provider for the financial services industry. While our company is best known for our hosted risk-based authentication and payment security services, we provide solutions ranging from mobile payments, fraud and risk management, prepaid solutions, and a host of merchant services.
Wibmo, based in Cupertino, CA, has built state-of-art, large scale, and secure operational capability that processed over a billion transactions last year. With a significant footprint in APAC, Wibmo is growing our presence globally, while maintaining a strong presence in the region.
Tackling the Issues That Reduce Speed of Delivery
As Wibmo's director of IT, my primary responsibility is to build, maintain and manage the production systems and ensure IT Infrastructure is scalable and robust to support continuous availability and sustain growth. My team provides the capacity and expertise our business development people need to grow the company and to provide high-quality service and products to customers in the banking and ecommerce sectors.
One of the crucial determinants of Wibmo's success or failure is our speed of delivery. When I joined the company two years ago, it was not one of our strengths. Our production environment was running on physical servers, and every time we added a new customer, we had to order and install new hardware. This slowed both our sales and development process.
When our sales team sat down with a client, the best they could offer was an eight- to ten-week turnaround, in which four- to five-weeks was for infrastructure procurement. They would plan the specs and requirements, lay out the functionalities, and then turn to our internal technology teams and ask them to make everything work. The process was convoluted at the best of times.
After our finance team had reviewed the initial paperwork, our IT department performed due diligence to establish the procurements required to meet the specs. Then, vendor evaluation was carried out which had its own timelines before we zeroed-in on a vendor. After multiple iterations, the purchase order was approved and issued, which would take significant time and effort for various teams.
In most cases, it took five or six weeks for us to receive the hardware, and then two more to install and configure it.
A few months after I started working at Wibmo, speed became our new mantra, and that meant upgrading our infrastructure. For the first 16 years of our existence, we'd run everything on bare-metal servers. When the subject of virtualization came up, no one was ready to risk it. Everyone thought that dedicated processors and dedicated drives were the right answer. Nobody thought that virtual machines, let alone software-defined IT infrastructure, could solve our problems.
Limiting our Proof of Concept to Counter Resistance
To counter this internal resistance, I decided to proceed with a limited trial. We started with some midrange HPE DL 360 rack servers. They had average compute capabilities, didn't break the bank, and ran VMware well enough for our proof of concept.
The results were impressive. Instead of spending five to six weeks to commission a single server for one of our customers, we could set up a virtual server with all the necessary applications and customizations in just a few hours.
Once we'd proven we could deliver solutions using this simplest of setups, we began to add more rack servers and virtual machines. However, we quickly discovered that traditional hardware was not capable enough to sustain our growth and came with its own set of problems.
Old-school rack servers took up too much space in our data center and needed more power to run and keep cool. Connecting compute and storage modules also required a mess of copper and fiber optic cables. Although the ability to host 10 to 25 virtual machines per box was a step in the right direction, a traditional server farm was not very scalable. Despite the lower initial investment, they cost more to operate in the long term.
It was time to take a deep dive into software-defined infrastructure. We'd successfully made a case for virtualization, and were now ready to push the technology to its limits.
Virtualizing More to Reduce Our Carbon Footprint and Our Expenses
Since our entire infrastructure runs on HPE hardware, we stuck with the company as our vendor of choice. We began replacing some of our older servers with HPE Synergy 1200 frames, each of which can house up to 12 dual-processor server blades. We can drop in blades as needed, and there is only one set of LAN and fiber optic cables connecting each of the frames to our switches and our storage arrays. A single frame running virtual machines can replace dozens of bare-metal servers, thus massively reducing capacity.
As expected, moving to HPE Synergy reduced our carbon footprint and our capital expenditures, but it also allowed us to lower costs and increase flexibility for our customers. In the past, we had to lock in the number of servers allocated to a bank at the start of a service agreement, and then add physical infrastructure as needed. This is no longer the case.
If we need to increase or lower customer capacity now, we can allocate and balance resources on a single HPE Synergy frame. We no longer have to spend weeks procuring and configuring individual servers. New accounts are up and running in a couple of hours, and existing customers can add or subtract resources as needed.
Once we started moving our server architecture to Synergy frames, it became clear that our storage setup wasn't fast or flexible enough to keep up with our software-defined architecture. We added HPE 3PAR flash arrays to our existing two- and four-terabyte SATA and SAS hard drives.
With the adoption of HPE 3PAR storage, we moved to a multi-tier architecture that allows us to segregate application processing and data management. For example, we can now host a high-traffic database to one of our newer flash drives, and a low-IOPS archive on an older SAS hard drive. Again, this has a tremendous impact on efficiency and our bottom line because we can allocate customer resources without having to buy new hardware every time they ask for additional capacity.
Using a Single Pane of Glass to See More of Our Infrastructure
Along with our new infrastructure, we also adopted HPE InfoSight as our server management and support platform. InfoSight provides a single pane of glass management view. It employs AI and predictive analytics to monitor our compute and storage environments in real time.
HPE InfoSight's Global Operation and Wellness Dashboards offer a consolidated view of the status, performance, and health of server infrastructure. InfoSight can also isolate performance bottlenecks and recommend ways to overcome them.
A few weeks ago, one of our application teams contacted us because one of their servers was experiencing latency issues that were causing a significant dip in their app's performance.
We had a joint meeting, and we used HPE InfoSight to show them how long it took that particular server to respond to a request from their web server, and how long it then took to forward that request to the database. We were able to pinpoint the location and the duration of the bottleneck in real time. It was only a few milliseconds, but it was significant.
Once we'd determined that the slowdown was not the result of an infrastructure issue, the team was able to go back and isolate the patch that was causing the problem. We could not have provided this level of detail using our previous server management tools.
Adopting Flexible Capacity Services to Further Reduce Our Capital Expenditures
The final element in the transformation of our IT infrastructure is HPE Greenlake Flexible Capacity Services. Greenlake provides consumption-based, metered IT services onsite rather than in the cloud, and allows us to better manage our capital expenditures.
With HPE Greenlake, we don't pay upfront for our IT infrastructure, but only for the capacity we use. Because we are not buying our equipment outright, our expenses are lower, and we can add and pay for servers and storage space on an ad hoc basis.
Thanks to HPE Greenlake's flexible pricing and the efficiencies of HPE Synergy and HPE 3PAR, we have reduced our IT capital expenditures by 80% while at the same time dropping our speed to delivery from weeks to hours. Customer service and satisfaction are now at all-time highs, and our infrastructure has never been better optimized.
In many ways, the members of our finance team have benefited most from these improvements. They no longer have to spend their time getting quotes from multiple vendors, and they no longer need a dedicated person to handle purchase orders.
Starting Small and Building on Modest Victories
Our IT staff is now more efficient and better motivated. The HPE solutions we have adopted allow us to do more with the time we have. Our monitoring, management, and support activities have been streamlined, freeing us to take additional training that will further improve our service levels.
Internally, we'll be enhancing Wimbo's security controls and back-up solutions. Our reduced capital expenditures and our improved infrastructure have empowered my team to start looking at the technologies that will define the future of this company.
I'll leave you with this final thought. Go slow. Starting small is the best way to approach the future. One modest victory leads to another. Before you know it, everything will have changed.
Our executive and finance teams are very conservative. They would never have accepted the wholesale replacement of our physical IT tools with virtual machines and software-defined infrastructure, but we validated our ideas by succeeding at one tiny thing after another. Taking your time is the best way to ensure that everything works and everybody wins.