Following the sale of assets, Client was required to move out of its existing data centers and migrate its applications and data to new facilities. Metagyre was contracted to plan, design, build out and bring Client on-line in their two new collocation centers. The new site in Reston, Virginia replaced the services provided by the Boston data center and a Tukwilla, Washington location replaced the Seattle data center.

This project consisted of building out two new collocation centers, deploying network, server and storage equipment followed by migrating services and bringing them on-line. Client’s business required that the equipment, applications and data move without disrupting their 24×7 service.

Client uses a number of technologies to deliver world class metasearch services to their customers. These include: NetApp storage, Cisco, Checkpoint, Netscaler, HP Storage, HP Blades, Netezza, VMware, Linux, MS Windows 200X, MS SQL, Oracle, MS Exchange, Active Directory, SAP.

Design and Build Out

• Aid in the selection of two collocation facilities.
• Design and build out the new facilities including structured cabling, electrical distribution and internet connectivity.
• Design rack layouts, server placement and port assignments.
• Manage the procurement and deployment of network, storage and server equipment into the new data center.
• Schedule and manage all professional services and sub-contracts for activities requiring third party activities or equipment purchases. Professional services included the configuration of new networks, storage assignments and virtualization services.
• Plan application and data move tasks for all applications within the two data centers including the test and roll back plans for each.
• Secure stakeholder buy off for each of the applications (owned or shared by multiple stake holders).
• Test core data center infrastructure, redundant power and networking.
• Stage, test and verify networks, servers and storage as operational.

System Migration
Mission critical application data was replicated to the new facility and kept in synchronization until the system was cut over. The application servers (95% virtualized at 1300 servers), were migrated using VMware capabilities in conjunction with their supporting data and shared systems (SQL Server, Oracle, Exchange, LDAP, FTP, DNS). Because of the complexities and interrelationships of the applications and data, this move required a complete data center cut over. For each data center (applications, data, middleware, shared systems) the applications were migrated electronically over the network, brought up, synchronized, tested and run in parallel before bringing the data center into production. In order to remove risk, as each new data center was brought online, traffic loads were managed and monitored to ensure quality. Load steps were set at 5%, 20%, 50%, 75% and 100% with mitigation and go/no go decisions established for each one.

Following the migration of the two data centers’ client facing production services, the back office and reporting systems were brought on line allowing Client to continue to provide world class metasearch capabilities without experiencing any customer down time.

Obstacles
After scheduling out the two data center build outs and migrations, it became evident the time line exceeded Client’s contractual commitment to vacate their current data centers. Metagyre proposed rescheduling activities to build out both data centers concurrently. Client accepted the recommendation to their original plan to build out the data centers one at a time. The dual data center build out approach allowed Client to complete the project ahead of their vacate commitments.

Project Summary
Client’s data centers were successfully relocated in less than the originally budgeted amount and time frame. By optimizing activities across the two data centers in parallel, Metagyre was able to reduce the time line without adding additional risk to the project. Following the move, Client has continued to expand its online metasearch services by taking advantage of the physical and virtual service growth plan developed by Metagyre.

Need help with your data center infrastructure? Contact theData Center Solutions experts at Metagyre, Inc. Data center management to data center consolidation putting business value in the data center.

Dual Data Center Go Live

At a high level there are three areas of guidance for a data center migration. Those areas are Data Center Move, Data Center Design and ongoing Data Center Management. Although their are some universals ideals, each company has to develop their own guiding principles. Individualized guiding principles are required since they reflect corporate culture, technology strategy and business goals. For a good start on developing your own data center migration guiding principles consider the following:

Data Center Move Principles:

1. Business Drivers:
   1. Contract renewals are driving the completion date Production cut over must be on or before the current contract expiration date
   2. Down time is expensive and taken as a last resort
   3. The data is the asset and needs to be protected
   4. One-time use equipment or service should be minimized
2. A team approach will be used in the move so that all disciplines are represented on the team.
   1. Storage: design, document, perform and sign-off on all storage component activities. Include vendor management for physical move of the storage component
   2. Networking: design, document perform and sign-off on all network component activities
   3. Systems: design, document, perform and sign-off on all systems component activities
   4. Back office: design, document, perform and sign-off on all back office component activities.
   5. Application Development: provide information, documentation, guidance and sign-off on all application component activities.
   6. Software QA: Test and sign-off on all functionality
   7. NOC (network operations center): Gatekeeper for documentation, escalation and monitoring of move activities
   8. Business Owner: Sign-off and approval of all move activities
3. Extend the network so both spaces look like one
4. Extend the SAN so storage in both spaces looks like one. Vendors will be enlisted in moving and ensuring proper turn up of their storage systems. New storage equipment will be used as a temporary landing zone for data in order to move it off of a subsystem prior to its move and then back onto it after its move
5. Pools of resources will be split and moved in stages. One resource should not need to know the physical location of another resource
6. Designs will be published, reviewed, challenged and revised prior to the move. During the move designs and scripts will be followed

Data Center Design Principles:

1. Plan Ahead:
   1. Data center plans and designs are to support the next three (3) years of anticipated growth.  This growth is estimated at 20% year over year
   2. Environmental monitoring is required in unmanned facilities
   3. Equipment will be maintained on a five (5) year life cycle
2. Keep it Simple:
   1. Aggregate cabling where possible to segregate patching
   2. Standardize on the rack, ladder system, PDUs, Patch Panels and cabinet layout.
   3. Priority deployment of standard computing:
      1. Virtual server on x86 architecture
      2. Blade Center
      3. Standardized x86 based server platform
      4. Other architecture
3. Remain Flexible
   1. Equipment can be placed in the next available cabinet.  There is no intelligence built into the floor plan
   2. New standards should be carried forward and retrofitted only where it makes business sense
4. Think Modular:
   1. Two cabinets make a non-blade center module to be deployed when additional resources are required
   2. One cabinet makes a blade center module to be deployed when additional resources are required
   3. Cabinet modules are outfitted the same for network and cabling
5. Keep things covered, bundled and out of sight:
   1. Cabinets have doors and sidewalls
   2. Copper cables are patched within the cabinets using the provided cable management
   3. Cables follow standard paths and do not cut across cabinets
   4. Fiber and copper cables are kept in separate trays
   5. Blanking panels will be used

Data Center Management Principles:

1. Label Everything:
   1. Racks, PDUs, and environmental monitoring devices are labeled with name as it appears in the book of record
   2. Servers are labeled with name as it appears in the book of record
   3. Network equipment and patch panels are labeled with name as it appears in the book of record
   4. Storage is labeled with name as it appears in the book of record
   5. Fiber, copper and power cables have a matching serialized label at both terminating ends. (should track in the book of record what cable attached to equipment)
2. Configuration Management is paramount:  the book of record contains information on all physical and virtual servers, network equipment, storage systems, racks, PDUs and environmental devices
3. Change Control:
   1. The change approval board ensures the inventory is updated
   2. All work requires a change control
4. Access is Locked Down:
   1. Only 3-5 people have physical access to the data center
   2. The Facilities group manages access to the data center
   3. Change control approval is required to perform work in the data center
   4. The NOC is aware of data center access through the daily production schedule and notices
5. Audit the work: Audits are conducted on
   1. The book of record against the physical data center
   2. Data center sign-in against daily production schedule
   3. The NOC is the Gate Keeper for data center management process adherence

These data center migration principles valid when moving into or out of a collocation facility or an in-house data center.

Need help with your data center infrastructure? Contact theData Center Solutions experts at Metagyre, Inc. Data center management to data center consolidation putting business value in the data center.

Data Center Migration Principles