Modern enterprise operations are centered around data centers. Businesses, partners, and customers worldwide rely on the data center to deliver resources and services.

Small and mid-sized businesses can often build a proper “data center” in a closet or other convenient room with few modifications. However, the sheer scale of enterprise computing necessitates an ample, dedicated space designed to handle the IT infrastructure’s space, power, cooling, management, reliability, and security needs.

Regarding capital investment and recurring operational expenses, data centers represent the business’s largest and most expensive asset. Throughout the facility’s lifecycle and with changing business circumstances, business and IT leaders should pay close attention to the issues involved in data center design and construction.

1. Location and Site Selection:

Choosing the right location for a data center is crucial. Factors such as proximity to power sources, access to fiber optic networks, and environmental considerations must be considered. Additionally, site security, including physical security measures and disaster recovery plans, should be carefully evaluated during the site selection.

2. Infrastructure and Power:

Data centers require a robust infrastructure to support their operations. This includes redundant power sources, backup generators, uninterruptible power supply (UPS) systems, and efficient cooling mechanisms. Implementing energy-efficient technologies can help reduce operating costs and minimize environmental impact.

3. Network Connectivity and WAN Design Considerations

Reliable network connectivity is essential for data centers. Businesses should consider multiple internet service providers (ISPs) to ensure redundancy and minimize the risk of network downtime. Implementing high-speed, low-latency connections is crucial for meeting the growing demands of data-intensive applications.

4. Scalability and Flexibility:

Data center design should allow scalability and flexibility to accommodate future growth and changing business needs. Modular designs and flexible rack layouts make adding or removing equipment as required easier. Additionally, utilizing virtualization technologies can help optimize resource utilization and improve overall efficiency.

5. Security and Access Control:

Data centers house sensitive and valuable information, making security a top priority. Implementing robust physical security measures such as biometric access controls, video surveillance, and fire suppression systems is essential. Regular audits and assessments should be conducted to ensure compliance with security standards and industry regulations.

6. Environmental Considerations:

Data centers consume significant amounts of energy and generate heat. Designing energy-efficient data centers can help reduce operational costs and minimize carbon footprint. Innovative approaches, such as utilizing renewable energy sources and implementing advanced cooling techniques, can contribute to a greener and more sustainable data center infrastructure.

7. Monitoring and Management:

Efficient monitoring and management systems are crucial for maintaining optimal data center performance. Implementing comprehensive monitoring tools can provide real-time insights into power consumption, temperature levels, and network performance. Additionally, automated management systems can help streamline operations and minimize human errors.

For pre-information, you may find the following posts helpful.

1. Virtual Data Center Design
2. Modular Building Blocks
3. SDN Router
4. Internet Locator
5. Triangular Routing

Back to Basics with Data Center Design Guide

• A primary tenant container is a two-tier tenant model that contains a “public” zone and a firewall-protected “private” zone. Load Balancing services are available within each zone.
• Virtual Routing and Forwarding instances ( VRF ) contain tenant-routing information that is exchanged via OSPF.
• Tenant VLANs in the Layer 2-domain map to corresponding VRF. VLANs and VRFs provide path and device isolation.

This example utilizes Nexus 7000 series as the Aggregation Layer device. The services layer utilizes a Dedicated Data Center Services Layer ( DSN ) with Catalyst 6500 Virtual Switching System ( VSS ) instead of an external firewall and load balancing appliances. VSS provides redundancy and increased throughput, resulting in one control plane and two data paths. Chassis-based DSN layer provides firewall and load balancing services with Application Control Engine Modules ( ACE ) and Firewall Services Modules ( FWSM ).

• A key point: Routed and transparent mode.

In routed mode, FWSM acts as a router hop in the network. The routed mode supports many interfaces, and each interface can have its subnet. In transparent mode, FWSM acts like a bump in the wire ( Layer 2 firewall ), not a router hop in the network. FWSM connects the same network on its inside and outside interface. Set independent modes for each context’s mode to either routed or transparent mode.

 Key Data Center Design Points:

• OSPF for each Tenant. Essentially, two separate routing domains are connected via static routes. The first routing domain is the unprotected network and the second routing domain is the protected network. These domains don’t directly connect because their OSPF Areas don’t touch.
• Stub Areas and Not-so-stubby ( NSSAs) were chosen for the Stub area type. NSSA areas provide the benefits of a stub area and the capability to import external information, e.g., static routes.
• The “unprotected” VRF connects to OSPF Area 0.
• The “protected” VRF does not connect to any other OSPF Area. Static routes provide complete isolation and full reachability to the public zone. Static routes redistribute into OSPF at Autonomous System Border Routers ( ASBRs )

Tenant routing with FWSM in transparent mode

 Key Points:

• OSPF utilized for each Tenant results in a single routing domain.
• Stub Areas utilized and Not-so-stubby ( NSSAs) chosen for Stub area type.
• The “unprotected” and “protected” VRFs share the same OSPF NSSA area. OSPF NSSA extends through FWSM, creating one routing domain instead of two separate ones, as evident when FWSM acts in routed mode.

Following the two-tier model’s basic structure, one can scale out and build other tenant models, for example, a two-tier with a single firewall and multiple private zones, a two-tier with various firewalls and multiple private zones, a three-tier model, etc.