May 16, 2017

Software-Defined Storage Modernization Done Right:  [Case Study]


Optimization and Modernization Using On-Premises Storage-as-a-Service 

Industry: Hospitality 
Company Type: Private >325MM
Company Size: 5001-10,000 Employees
Location: Wisconson, USA

Services Performed:

  • Storage Assessment
  • Architectural Design Consulting
  • Storage-as-a-Service
  • Software-Defined Storage
  • Solution Implementation
  • Ongoing Storage Advisory Consulting


Like many organizations today, the IT department at this resort hotel chain is undertaking initiatives in an effort to lower the total costs of data center storage operations, modernize its storage environment and optimize the organization’s ability to deliver IT services that drive positive business outcomes for key internal stakeholders.  All of this all while ensuring the highest levels of data security and compliance.

In an attempt to do so, they have put off purchases of new enterprise storage hardware and continued using legacy storage arrays past their vendor end-of-life windows. While delaying upgrades did prevent capital expenditures "CapEx" purchases, the legacy arrays also caused numerous problems for the organization.

Business Drivers for Initiative:

  • Capacity Planning
  • Application Centric Storage Architecture
  • Enhance User & Application Performance
  • Storage Elasticity - Capacity and Performance 
  • Cost Containment & Cash Flow Predictability

The client decided it was time to conduct a thorough independent analysis and assessment of current and future storage capacity, usage, architectural and performance requirements.

Project Overview:

As with all storage assessments Headwaters Group conducted interviews with key stakeholders and gathered a meta-data from the existing storage environment.  An analysis of the data identified several key findings:

Existing Environment:

  • 3PAR 7200 disk arrays
  • NetApp FAS2050 arrays
  • NetApp FAS2240 arrays
  • IBM DS3400 arrays
  • Exagrid 4000 backup target
  • Exagrid 13000 backup target
  • Exagrid 40000 backup target
  • HP StorageServer
  • Sonicwall

Poor performance and outages

  • One legacy array (using slow spinning disks) was misconfigured by the vendor with a short stripe. On a regular basis, storage latency would increase to the point that VMs on the storage would go offline, causing havoc for application teams. Latency of hundreds of ms was common, and occasionally even over 1000ms
  • Infrastructure team personnel spent up to 50% of their time on these and other outages and problems

Lack of support and maintenance on arrays

  • It’s very common with older arrays to find that the firmware has not been kept up to date: this was very true with this client where nearly all arrays had downrev firmware
  • As with many organizations, the legacy arrays were under support contracts with 3rd-party vendors. These vendors supply hardware fixes only, and more complicated problems are extremely difficult to troubleshoot and resolve
  • There were four different array manufacturers, and almost all arrays in use were past end of vendor support
  • The small storage team just could not keep up with all the work of maintaining the heterogeneous arrays, and no one on the team had training or extensive experience on all the array types
  • The legacy arrays were a TICKING TIME BOMB!

Multiple protocols and point solutions for storage

  • In the past, it was common to deploy different arrays as point solutions for individual project requirements, which explained the clients’ 12 separate arrays spanning four storage technologies
  • Some arrays were used for block solutions, while others were for file, and different arrays served different performance requirements
  • Due to so many arrays and the small storage team, the storage environment was UNMANAGEABLE

Headwaters Group Recommendations

To solve storage problems the customer was experiencing, we knew that the future state storage architectures needed to have specific features and functionality:

  • Flash (SSD) and SATA Storage to accommodate the diverse performance requirements for multiple business applications
  • Elasticity in capacity and performance based on “just in time” business needs
  • On-Premesis and multi-cloud functionality for enhanced data security and  mobility
  • Consumption-based OpEx cost model that allows paying for only what they used
  • Software-defined storage that enabled multiple protocols (block, file, and object) to be delivered from the same storage device
  • Simple to manage. The small storage team needed a solution to the problem of ongoing storage management service, upgrades and maintenance for the storage environment, which they never had time for in the past

New Solution:
We recommended software-defined storage-as-a-service from Zadara Storage for this client, which has been a great fit for their needs. The customer deployed their first Zadara OPaaS (On-Premise as a Serviceunit around the end of 2016 and has been so happy they already ordered and received delivery of their second Zadara OPaaS unit. 

The OPaaS includes the following:

  • On-Premesis hardware and software system
  • All flash drives for extremely fast performance
  • Enough storage to satisfy the capacity needs of legacy arrays to be decommissioned
  • Additional drives (at no charge) to handle unexpected storage increases, client pays for these drives only as they are put into service
  • Multi-Cloud (Pulic and private)  replication and agility for sychronous data mobility
  • Virtual Private Storage Array (VPSA) technology that enables a multi-tenant capability with single tenant experiece 
  • 2 x 40Gbps switches for iSCSI access (FC also available)
  • 1 x 1Gbps management switch
  • 1 x 1Gbps firewall for management switch
  • Simplified data migration

Download The Zadara Enterprise Storage-as-a-Service eBook

Pricing with Zadara is entirely “pay-as-you-go” with no up-front cost. The customer chose a three-year commitment to benefit from the 50% price reduction. This customer is using only iSCSI/block storage so far but plans to use the built-in NAS (NFS/CIFS) on the Zadara array in future. They could also configure the same Zadara with Object storage if desired

Project Results:

The project was a success as the Client within 60 days was elated able to provide his business lines with drasticly improved performance while reduceing RTO down from days to 4-6 hours.

  • Complete elimination of outages (previously there were 1-2 outages per week due to problematic legacy array performance)
  • Complete elimination of user complaints about performance (frequent on old arrays)
  • Major reduction in fire-fighting time due to legacy arrays
  • Elimination of 3rd-party maintenance cost on old arrays 
  • Simplification of the storage environment
  • Huge reduction in day to day management and training requirements for storage team
  • Elimination of traditional cost associated with day-today storage maintenance and administration (admin and firmware updates handled by Zadara and with no downtime)
  • Data protection with sychronous replication 
  • End of need to forecast future storage requirements due to “pay-as-you-go” pricing of Zadara
  • Storage hardware refreshes provided free of charge every 39 months or less
  • Cloud transition path as Zadara can also live in OR replicate to Amazon, Azure, Google and 20+ other public cloud provides as well as private clouds.

Are you considering storage moderization?  Headwaters Group  can help - Learn More

Filed Under: Case Studies, Software-Defined Storage, Storage-as-a-Service

Latest News

What Our Customers Say

Concierge Practice Managers

“Headwaters Group is my go to services partner and my first choice in nearly every situation.” Insight Practice Director