Project Type: EMC/HP Data Migration, Enterprise Storage
|Industry:||Financial Services, Investment Banking, & Asset Management|
|Company Type:||Publicly Traded, Revenue > $550 Million|
|Company Size:||Approx 47,000 Employees|
|Location:||New York, New York – United States Headquarters|
- Data Migration Planning and Design
- Automation and Scripting
- Data Migration
- Data Replication
Our Client operates their U.S. Headquarters out of New York, New York, supported by four data centers. Three of these data centers are located in northern New Jersey with a fourth located in Manhattan. As a prominent international financial services company, the Client had decided to embark on a data center consolidation project. The consolidation project consisted of migrating 900+ servers from several older data center sites to a newer, larger site. Our client had decided to divest one of their legacy data centers and was required to vacate the premises by the end of October. Failure to do so would result in a multi-million dollar penalty, thus creating a large incentive to complete the migration ahead of schedule.
The project became increasingly complicated by the requirement to maintain service level agreements regarding system downtime which limited such outages to only 6 hours for most applications.
The data would need to be migrated by the beginning of September to ensure all remaining cleanup tasks could also be completed after the migration. Thus, ensuring the Client would have the sufficient time to vacate their legacy data center prior to the end of October. In February, the scope of the project changed to include the migration of both Tier-1 data (live production) and Tier-2 data (non-critical data, backup data, archived files) from EMC and HP to Hitachi.
The project consisted of two remote copy migrations performed in parallel and required a high-level of expertise in the areas of EMC and HDS array-based data replication. A total of 900+ nodes were identified to be migrated consisting of Tier-1 data (live production) and Tier-2 data (non-critical data, backup data, archived files) from EMC and HP to Hitachi Data Systems VSP arrays.
As a result of these time constraints, our engineer used script programming and the automation of manual tasks to reduce the time required to complete the set-up tasks and data replication processes. These scripts automated many of the time consuming storage administration tasks used to administer HDS, Brocade, Cisco, EMC, and HP equipment.
Headwaters offered additional value to the project team by supplying the capability to fast-track the migration process through advanced scripting and automation. Our engineer leveraged his development background of more than fourteen years of experience in writing scripts for both Headwaters and others.
In a typical project of this magnitude, auditing and planning the migration could could take approximately one to two weeks per server if done manually; however, script automation based off the utilized devices, sizes, HDS RAIDCOM and EMC INQ commands, zoning templates, and target device input made it possible for Headwaters to drastically cut down the time of such a process. Scripting and automation allowed our engineers to reduce the time required for the migration/replication process by approximately 75%.
As an example, the replication setup for 300 new SAN devices of various sizes typically takes three to four hours to complete, but with automation scripts, our engineers reduced this time to approximately 15 minutes. Whereas in years past, a migration/replication event that averaged six hosts per week was deemed acceptable, we found that if leveraging an approach based on automating many of the tedious, time consuming steps involved in planning and executing a migration, our engineers boosted this number to between 20 and 50 migrations per week.
- DMX 3, 4, 1000, and 2000
- SRDF – Open Replicator
- HUR (Hitachi Universal Replicator) – Asynchronous Replication
- Cisco Switches
- Windows Hosts
- Oracle Rack
- HP 3PAR
Products In Use:
- Mandated Exit of legacy Data Center by October
- SLA Time Constraints of 6-Hour Outage Window
- Increased Data Center Costs
- Outdated Hardware
- Possibility of an financial penalty for Not Meeting Set Time frame
- New Data Center
- All New Equipment Other Than Servers
- Storage Equipment refresh to HDS VSP for Tier1
- Moved From Thick Provisioning to Thin Provisioning
- Approximately 100 Terabytes in Capacity Savings
- Significant Performance Increases