Following its separation from eBay, PayPal needed to untangle nearly 20 years of shared infrastructure data and build its own data center management system. By implementing a new a Data Center Infrastructure Management (DCIM) platform powered by real-time data from AVEVA™ PI System™ and AVEVA™ PI Vision™, PayPal gained full visibility into power, space, and cooling at its data centers. The platform allowed PayPal to forecast budget accurately, streamline operations, and make faster, data-driven decisions across 33 data centers worldwide.
Challenges
Untangle 20 years’ worth of data from eBay’s AVEVA PI System to create an independent system
Needed to enable trusted data sharing between the two separate companies
Lack of real-time information inhibited critical business decisions
Results
Gained data-driven insight into power and cooling status of assets in data centers
Ability to forecast budget and predict future capacity needs
Reduced alert fatigue and automated service ticketing
Delivered executive-level reporting in seconds instead of weeks
Positioned to reduce co-location costs through better space and power allocation
Breaking up from a partner company is never easy, especially when all of your data is intertwined. First established in 1998, the mobile payment company PayPal has been a subsidiary of eBay for the majority of the last 20 years. PayPal now operates across 33 data centers and facilitates more than 6 billion payment transactions per year. In 2014, eBay announced plans to separate PayPal and turn it into its own company, initiating a long journey for PayPal toward independence. Shawn Tugwell, Senior Data Center Engineering Manager at PayPal, recently told the story of how PayPal disentangled its data from eBay’s and created its own data center infrastructure management (DCIM) tools.
Solution
Deployed an independent DCIM platform built around ServiceNow and powered by AVEVA PI System and AVEVA PI Vision to centralize data, automate alerts, and enable real-time infrastructure monitoring and forecasting.
All tangled up
eBay has long used AVEVA PI System to help monitor and manage its racks and data centers. As a subsidiary of eBay, PayPal had nearly 20 years’ worth of data wrapped up in eBay’s AVEVA PI System. The first step was for PayPal to create its own AVEVA PI System and migrate millions of data tags and other information related to asset framework and notifications in AVEVA™ PI Server. AVEVA PI System allowed PayPal to continue sharing information with eBay, as both companies still had assets in each other’s data centers, even after the separation.
PayPal also faced other challenges as it became independent. The company used a large number of spreadsheets and manual processes to track critical data center information on space, power, and cooling. These processes, combined with a lack of real-time data, prevented PayPal managers from making informed decisions about capacity planning. Fragmented monitoring, multiple platforms, and alarming systems produced a disconnected view of operations in the data centers.
PayPal wanted to bring its equipment usage and maintenance forecasting and planning, as well as data center visualization into one system. The goal was simple: enable data-driven decision-making using real-time environment information and alerting. Engineers also needed space and power visualizations to better plan their installation and decommission activities. PayPal enlisted the help of Casne Engineering and Deloitte Consulting to help with its data reorganization and to set up its own independent DCIM tools.
A data pipeline for a new DCIM platform
Once PayPal implemented AVEVA PI System independent of eBay, it became the data pipeline for the new DCIM platform. PayPal built the DCIM around ServiceNow, PayPal’s central platform for ticketing and workflow, as well as Mobile Reach and Tier 44. The DCIM uses AVEVA PI System for gathering data, alerting, and monitoring the environment in the data centers.
Data comes into AVEVA PI System from several sources, including PayPal’s electrical power management systems and building automation systems, and the Raritan power strips in the data centers. PayPal uses AVEVA™ PI Vision™ to create dashboards and the asset framework in AVEVA PI Server to organize the information coming out of the various systems, providing data modeling for equipment. Logistics teams can then validate requests for space and power based on real-time data provided by AVEVA PI System. “Now we are able to forecast out; we are able to control what comes into the data center,” said Shawn Tugwell, Senior Data Center Engineering Manager at PayPal. “We’re not having to run around to try to find new space because now we have a tool that’s helping us make that business decision.”
“We are doing this to provide better value, to enable the business to make better decisions based on data, not on guessing what is the next best course to install cabinets or to track power utilization.”
--Shawn Tugwell, Senior Data Center Engineering Manager, PayPal
AVEVA PI System empowers PayPal’s new DCIM with real-time data for monitoring all its data centers.
Machine learning in the data cockpit
In addition to collecting information from various data streams, PayPal also uses AVEVA PI System to manage its maintenance operations. Before the new DCIM was in place, PayPal was inundated with alarms. To address this issue, the company created a system to generate service tickets in ServiceNow based on AVEVA PI Server notifications. PayPal can now track alerts, trend alert data, and use the data for lifecycle management of equipment in its data centers.
AVEVA PI System also helped PayPal deliver real-time data to upper management. “The best part of it is, I can now provide this data to our upper management in seconds instead of 40 days later,” Tugwell said.
In the future, PayPal hopes to use AVEVA PI System data on space and power utilization to drive down costs with its co-location providers. PayPal has assets in these data centers that it doesn’t own. The company plans to work with a hardware engineering team to better allocate the power in those data centers. “That is going to save us money because we won’t need to buy more co-location space,” Tugwell said.
Product highlights
AVEVA™ PI System™
Collect, aggregate, and enrich real-time operations data for immediate problem-solving and easily deliver formatted data to enterprise applications and advanced analytics.
AVEVA PI Server
Learn about the trusted, high-volume, and real-time data storage, contextualization, analytics, and notification engine at the heart of AVEVA™ PI System™.
AVEVA PI Vision
With AVEVA PI Vision, turn raw data into rich, visual displays and share valuable insights across your enterprise.