Top Banner
Technical Case Study This Is What Agility Looks Like How Revlon IT harnessed big data with a private cloud and global master data model If you visit Revlon, Inc.’s data center in Oxford, North Carolina, don’t blink, or you’ll miss the infrastructure. Just two racks house roughly 3.6PB and 800 virtual servers that process an average of 14,000 transactions per second (TPS) from systems around the world, with 99.9999% uptime. When people walk into our data center, they ask, “That’s it?” The answer is yes, and it runs everything. The picture was starkly different just a few years ago, when business intelligence was buried in 21 separate enterprise resource planning (ERP) systems. High management overhead kept our IT team from saying yes to requests for new applications. Buildings didn’t even have Wi-Fi access, so chemists had to jog back and forth between mixing vats and manufacturing control systems to adjust product formulations. This case study describes how we transformed Revlon IT by building a private cloud and putting in place a global master data structure. Along the way, we converted IT from an impediment to an enabler for agile strategic planning, product development, manufacturing, warehousing, distribution, and operations. Now systems work for people instead of people working for systems. We accomplished the transformation in three phases: building a private cloud, insourcing disaster recovery (DR), and building a global master data model. Phase 1: Building a Private Cloud We began the transformation effort by bulldozing the existing IT environment. This included re-IPing all devices and systems and building a private cloud. Our guiding principle for the cloud architecture was simplicity: single vendors for storage, servers, virtualization, and networking and a single Domain Name System/Dynamic Host Configuration Protocol (DNS/DHCP) server, server image, and desktop image. By David Giambruno, Senior Vice President and CIO; Ben Gent, Global Storage Engineer; and Michael Cannella, Data and Global Network Engineer
7

Revlon Technical Case Study

May 17, 2015

Download

Technology

NetApp

If you visit Revlon, Inc.’s data center in Oxford, North Carolina, don’t blink, or you’ll miss the infrastructure. Just two racks house roughly 3.6PB and 800 virtual servers that process an average of 14,000 transactions per second (TPS) from systems around the world, with 99.9999% uptime. When people walk into our data center, they ask, “That’s it?” The answer is yes, and it runs everything. Find out more about Revlon and NetApp here: http://www.netapp.com/us/campaigns/builton/?REF_SOURCE=smctwitter-initiative-builton
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Revlon Technical Case Study

Technical Case StudyThis Is What Agility Looks LikeHow Revlon IT harnessed big data with a private cloud and global master data model

If you visit Revlon, Inc.’s data center in Oxford, North Carolina, don’t blink, or you’ll miss the infrastructure. Just two racks house roughly 3.6PB and 800 virtual servers that process an average of 14,000 transactions per second (TPS) from systems around the world, with 99.9999% uptime. When people walk into our data center, they ask, “That’s it?” The answer is yes, and it runs everything.

The picture was starkly different just a few years ago, when business intelligence was buried in 21 separate enterprise resource planning (ERP) systems. High management overhead kept our IT team from saying yes to requests for new applications. Buildings didn’t even have Wi-Fi access, so chemists had to jog back and forth between mixing vats and manufacturing control systems to adjust product formulations.

This case study describes how we transformed Revlon IT by building a private cloud and putting in place a global master data structure. Along the way, we converted IT from an impediment to an enabler for agile strategic planning, product development, manufacturing, warehousing, distribution, and operations. Now systems work for people instead of people working for systems.

We accomplished the transformation in three phases: building a private cloud, insourcing disaster recovery (DR), and building a global master data model.

Phase 1: Building a Private CloudWe began the transformation effort by bulldozing the existing IT environment. This included re-IPing all devices and systems and building a private cloud. Our guiding principle for the cloud architecture was simplicity: single vendors for storage, servers, virtualization, and networking and a single Domain Name System/Dynamic Host Configuration Protocol (DNS/DHCP) server, server image, and desktop image.

By David Giambruno, Senior Vice President and CIO; Ben Gent, Global Storage Engineer; and Michael Cannella, Data and Global Network Engineer

Page 2: Revlon Technical Case Study

2

We selected NetApp as our storage and data management vendor, deploying NetApp FAS6000 and FAS3000 systems in the main data center and NetApp FAS2000 systems in four remote facilities around the world. Had we gone with a multivendor storage environment instead, we would have had to invest resources in making them play well together. As it is, we can spend that time developing new business intelligence applications.

After deploying the production infrastructure, we virtualized servers and desktops, beginning with low-risk applications such as print and file servers, moving on to departmental applications, and finishing up with mission-critical applications such as ERP and manufacturing resource planning (MRP). Today, more than 530 applications—97% of the total—and all desktops are files (virtual machines) running on VMware vSphere® and hosted on high-availability NetApp FAS6040 and FAS6240 systems. We’re running at 99.9999% uptime—fewer than 13 seconds of downtime annually. The private cloud processes 15,000 application moves monthly.

The NetApp systems also host SAP® front-end software; electronic data interchange (EDI) transactions; Microsoft® SharePoint® Portal Server; the mail transfer agent (MTA) used by IBM Lotus Domino; DNS/DHCP servers; time tracking; applications based on UNIX®, including SAP databases and bolt-ons; Oracle® E-Business Suite Financials; and dozens of other business-critical Oracle applications such as the Revlon manufacturing support and legal systems.

Server virtualization enabled us to consolidate from 160 to 70 servers, a 56% reduction. The smaller footprint reduced our global power consumption by 72%.

Redundancy doesn’t matter in our environment. We have only one server connection because if we lose a server, applications move to another server in the cloud.

T�el�e to IT Agility: 0-100% � Five YearsNo incremental spending to achieve results.

18 m�ths 3 years 5 years

InfrastructureSimplification

Global CloudDeployment

Global Cloudin Production

• Reduced costs with$70+ million in savingsand avoidance

• 425% more IT projects

• Increased businesscapabilities

Revlon Data Center Infrastructure Today• Simpler, smarter data management:

– 120% production storage utilization

– 1,500% disaster recovery storage utilization

– 1 person to manage 18 global NetApp® storage systems

• Flexible data scaling:

– 5 times more storage capacity at the same cost

– 14,000 transactions per second

• Highly available and reliable data:

– 99.9999% (six 9s) availability, with fewer than 13 seconds of downtime annually

– 15,000 application moves monthly

– 30TB of changes weekly flowing over the network to the DR data center

• Consolidation cost reduction:

– 56% server reduction, from 160 to 70

– More than $70 million in cost savings and cost avoidance in past two years

– 72% decrease in data center power consumption

• Agility:

– 70% faster application deployment

• Efficient project completion:

– 425% increase in the number of IT projects delivered

– 70% reduction in the time to deliver IT projects

– 99.6% of projects completed on budget and on time

Page 3: Revlon Technical Case Study

3

Phase 2: Insourced Disaster RecoveryNext, we increased the return on investment from NetApp technologies by bringing disaster recovery in house, an effort that paid for itself in just seven months.

Because servers are now files, we can quickly replicate every production server to a NetApp FAS3070 system located at our DR site, using NetApp SnapMirror® replication technology. We transmit 30TB of changes weekly to the DR data center.

In our four global factories, data centers have been replaced with preloaded “mini-me” data centers, which are half racks populated by multiple NetApp FAS2020 and FAS2050 systems and a commodity switch and server. We like to say we went from two racks of mess to one-half rack with 10 times the performance at one-eighth the cost. Like the production data center, the mini-me data centers replicate data to the DR site using NetApp SnapMirror software.

We test disaster recovery quarterly, bringing up an instance of all of Revlon—or an individual country, site, or application—in the DR data center. The solution was put to the test in 2011, when a fire destroyed most of the Venezuela facility. With our previous infrastructure, recovery would have cost millions of dollars and taken one to two weeks while people flew back and forth with suitcases full of tape backups. With the NetApp DR solution, in just one hour and forty-five minutes we brought up factory, EDI, and ordering systems in the DR data center and spun up virtual desktops for employees so that they could work from home or business partner offices the next day. We would have been back in business even sooner if our storage engineer had not been on vacation with his phone on silent.

We took the same steps to make sure Hurricane Sandy didn’t disrupt the business. As Sandy came barreling toward the Metro New York area in November 2012, we used NetApp SnapMirror replication technology to move our New York and New Jersey data center operations to the North Carolina data center—in one hour. After we confirmed the facilities had escaped damage, we moved the files back over the network to their original data centers.

Phase 3: Corralling Big Data with Master Data ModelAfter building the private cloud and DR infrastructure, we turned our attention to transforming big data—3.6PB worth—from a burden to a business asset. Revlon tracks 660 million SKU attributes monthly, and our job in IT is to convert this data to actionable information. Previously, requests for new business intelligence applications had to wait until we added storage capacity; created a mirrored copy of production databases; and performed extract, transform, load (ETL) operations—not exactly a recipe for agility.

We came up with a unique solution to the big data challenge by creating a global master data structure that automatically sorts, organizes, and structures all Revlon data originating anywhere in the world. Every 15 minutes, Revlon’s infrastructure pulls new data—including structured data as well as unstructured data—from all global systems. Microsoft Data Quality Services (DQS) cleanse and profile the data, normalizing it without the overhead of ETL.

After designing the master data model, we created more than 20 multiterabyte parent datasets. Now development and test engineers use NetApp FlexClone® technology to clone the datasets in an instant and run analysis in hours instead of weeks. The NetApp Data ONTAP® operating system creates these clones with zero storage overhead. Only new writes take up storage space, eliminating the costs ordinarily associated with database copies. After using clones, developers can either make them permanent or destroy them.

David GiambrunoSenior Vice President and CIO Revlon

David Giambruno joined Revlon as CIO in 2006. He is responsible for driving global technology capabilities to provide competitive advantages to the company’s business units and shareholders. The breadth and strategic nature of this project and the speed at which it was accomplished demonstrate his technical and business leadership. He and his team have transformed Revlon’s IT operations.

Page 4: Revlon Technical Case Study

4

The ease of creating zero-overhead clones is transformative for Revlon. There is no cost to mine the data and experiment. Instead, the NetApp systems are simply providing added value while they spin platters. We’ve used clones for large-scale application testing, change control, creating golden images, and even modeling the effect of different e-mail retention policies on storage requirements.

We think about big data in a different way than most companies. The traditional model has applications at the center, surrounded by big data that the IT team needs to ETL. In our model, big data sits in the center, with applications surrounding it (Figure 1).

ApplicationPortfolio

Current Big Data Practices Revlon Big Data Practices

Figure 1) Revlon big data practices.

Agility Means Saying Yes to the BusinessThe combination of the NetApp infrastructure, our global master data model, and mobility solutions lets us wield technology as a competitive advantage. We like to say we’ve taken the infrastructure out of the way so that the business doesn’t even notice it anymore. We can deliver our application portfolio to any device, anywhere in the world. And we’ve accelerated new application delivery by 70% on average.

The proof of this newfound agility is that Revlon IT delivered 425% more projects in 2012 than in 2007, reduced time to deliver a project by 70%, and completed 99.6% of projects on budget and on time. That’s our key statistic, and it’s largely the result of time savings from NetApp’s service automation and analytics tools. For example, waiting time to receive a server has dropped from 6 to 8 weeks to 15 minutes. Our business users never need to wait for servers, connectivity, or storage. If a department wants to try a new application, we generally have it up and ready before the meeting is over, and it costs us nothing.

The new agility ripples through Revlon:

• Supply chain optimization. The first application to take advantage of the master global data model was a global inventory report with drill-down capa-bilities, which distribution teams can access from tablets and smartphones. Around 80% of Revlon users just want key performance indicators (KPIs)—actionable information they need to do their job. Now they don’t need to take the time to log into SAP or Oracle for commonly needed metrics.

• Increased manufacturing efficiency. Instead of climbing up and down ladders to mixing tanks to check readings on wall-mounted computers, Revlon chemists use iPad® devices with custom apps to adjust settings on a formulation system hosted on NetApp storage. Not having to walk back to the office to adjust a formula avoids materials waste and manufacturing delays.

Ben GentGlobal Storage Engineer Revlon

Ben Gent has worked for Revlon since 2001. He held positions with the help desk and PC support before stepping into his role as global storage engineer, in which he is responsible for managing 3.6PB of data with 30TB of changes per week—a job that previously required 10 people.

Page 5: Revlon Technical Case Study

5

• Automated picking and packing. Rather than packing shipments based on a paper report that might be out of date, warehouse employees refer to a screen in their truck to see where to go, what to pick, and where to place the merchandise. Cutting down the time to package a product and get it out of the building increases the rate of inventory turnover. And when the workers in the truck can make accurate moves, we avoid vendor-imposed financial penalties for incorrect shipments.

• Mobility. The NetApp infrastructure is even helping to improve work-life balance. For example, finance teams working long hours during month-end close activities can access their virtual desktops on NetApp storage from home, using an iPad. And our IT team can change a server configuration or move data during the week instead of coming in on the weekend.

• Reinvented help desk. The IT team processed 239% more help desk tickets monthly in 2012 than we did in 2007, a benefit of spending less time on storage management. We resolve 85% of issues on the first call, a rare achievement. We’ve also seen a huge drop in the number of help desk tickets for problems and a correspondingly huge increase in requests from people wanting to get work done in a more efficient way.

In summary, the NetApp infrastructure and our master data model have trans-formed big data from a burden to a competitive advantage, even while cutting costs. Overall, IT has improved the bottom line for the business in the past two years through more than $70 million in cost savings and cost avoidance. In addition, we’ve practically eliminated the incremental costs and wait time to experiment and are limited only by our imagination.

Revlon employees, too, enjoy a new relationship with data. From anywhere, any time, they can connect to the information they need to do their jobs more efficiently and effectively.

How We Did ItNetApp storage infrastructure allowed us to disconnect innovation from incre-mental costs for networking, servers, storage, and staff time. Reasons include:

• We can commit resources without having a physical disk, a benefit of NetApp thin provisioning. Utilization rates for production and DR storage are 120% and 1,500%, respectively.

• Automated processes keep staffing requirements down. For example, NetApp OnCommand® Unified Manager provides a single view of storage management, provisioning, and protection. NetApp Workflow Automation creates a policy-based data infrastructure that we use to construct or edit workflows. We use NetApp OnCommand Insight software to plan capacity requirements and make sure that availability and performance meet service-level expectations. Before, we needed 10 people to manage storage. Today, people are amazed to hear that just one person manages 18 global storage systems with thousands of spindles.

• We lowered storage costs using NetApp Flash Pool™, which is part of NetApp Data ONTAP. Flash Pool lets us mix solid-state disks (SSD) and relatively inexpensive SATA drives within a single aggregate, or “hybrid aggregate.” Adding SSDs to the aggregate increased the performance of our busy SATA disks by a factor of ten, providing performance similar to serial-attached SCSI (SAS) disks at far less cost. We have been able to achieve a similar experience with NetApp Flash Cache™. We were able to defer buying new hardware until a budget cycle could be built and engineering work done to find the best solution. Configuration is basically zero effort. As of this writing, we’re also completing testing of NetApp Flash Accel™, which lets us use relatively inexpensive SATA disks by minimizing latency by up to 50%. This gives our scientists and marketers faster access to the data that drives product creation and sales.

Michael CannellaData and Global Network Engineer Revlon

Michael Cannella has been a data and global network engineer with Revlon since 2007. During his tenure, he has contributed to the design and implementation of the Revlon global private cloud and virtualization platform, an infrastructure that has transformed IT and helped align it to Revlon’s business.

Page 6: Revlon Technical Case Study

6

Our business users can continue to access data even during system maintenance because of the nonstop operations capability in the NetApp Data ONTAP operating system. We no longer need to schedule downtime for maintenance and caching because NetApp SnapMirror replication technology can move any virtual server, including ERP and other critical applications, between data centers. When we upgrade our server hardware, we can move virtual machines to the DR site while they continue to run and then move them back when the server upgrade is complete. Case in point: In 2012, we replaced Revlon’s global SAN core and data center server with no outages.

Another way we protect data is by using NetApp embedded security technologies, such as NetApp Storage Encryption, role-based access control (RBAC), and embedded antivirus.

Project Phases

PROGRAM STEPS

Built private cloud in internal data center • Built virtual server platform using NetApp storage systems

• Deployed SAN• Implemented wide area file services

Built global disaster recovery site • Replicated all sites to internal DR site in United States using NetApp SnapMirror replication technology

• Automated backup and recovery, 7TB weekly in just one building

• Tested quarterly

Shipped “mini-me” data centers (half racks containing NetApp FAS2000 series storage systems) to regional data centers

• Migrated application and desktop files remotely• Replicated data to DR site

Virtualized all resources: application servers, domain controllers, databases, file servers, and print servers

• Replicated everything rather than trying to understand all dependencies to selectively replicate

• Standardized images

What’s NextWe’re continually looking for ways to increase the value of our NetApp investment to the business, and some plans include:

• Simplifying the application portfolio: Our goal is to shrink the portfolio by 78% by 2014. As part of that effort we’ll collapse 21 ERP systems into one.

• Delivering marketplace applications that employees can download and use on iPhone® or iPad devices.

• Extending cloud access to partners and suppliers, giving them their own applications.

• Implementing an active-active architecture.

• Deploying NetApp Flash Accel to production environment, accelerating application performance so that product development and marketing teams can work faster and increase Revlon’s competitive edge.

Page 7: Revlon Technical Case Study

www.netapp.com

Follow us on:

© 2013 NetApp, Inc. All rights reserved. No portions of this document may be reproduced without prior written consent of NetApp, Inc. Specifications are subject to change without notice. NetApp, the NetApp logo, Go further, faster, Data ONTAP, Flash Accel, Flash Cache, Flash Pool, FlexClone, FlexVol, OnCommand, SnapManager, and SnapMirror are trademarks or registered trademarks of NetApp, Inc. in the United States and/or other countries. iPad and iPhone are registered trademarks of Apple Inc. Microsoft, SharePoint, and SQL Server are registered trademarks of Microsoft Corporation. Oracle is a registered trademark of Oracle Corporation. SAP is a registered

trademark of SAP AG. View, VMware, VMware vCloud Director, and VMware vSphere are registered trademarks of VMware, Inc. UNIX is a registered trademark of The Open Group. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such. NA-168-0213

About Revlon, Inc.Revlon is a global color cosmetics, hair color, beauty tools, fragrances, skincare, anti-perspirant deodorants and beauty care products company whose vision is Glamour, Excitement and Innovation through high-quality products at affordable prices. For more information, please visit www.revlon.com.

About NetAppNetApp creates innovative storage and data management solutions that deliver outstanding cost efficiency and accelerate business breakthroughs. Discover our passion for helping companies around the world go further, faster at www.netapp.com.

Go further, faster®

NetApp Products• NetApp FAS6000, FAS3000,

and FAS2000 storage systems

• NetApp Data ONTAP operating system

• NetApp Flash Cache

• NetApp Flash Pool

• NetApp FlexClone technology

• NetApp Operations Manager

• NetApp OnCommand management software

• NetApp SnapManager® for Oracle

• NetApp SnapManager for SQL Server®

• NetApp SnapManager for Virtual Infrastructure

• NetApp FlexVol® technology

• NetApp deduplication

• NetApp SnapMirror replication technology

• NetApp thin provisioning

Protocols• NetApp SAN (FC and iSCSI)

with CIFS and NFS

Third-Party Products• VMware® View® 5.1,

VMware vCloud Director®, VMware vSphere, and Site Recovery Manager

• IBM Lotus Domino mail

• Oracle E-Business Suite Financials

• SAP database

• Microsoft Exchange 2010 and SharePoint 2010

• Juniper switches

Another NetApp solution delivered by: