Cisco CCDA Cert Guide: Data Center Design

This chapter is from the book 

Foundation Topics

This chapter covers general enterprise data center considerations that you need to master for the CCDA exam. It starts with a discussion of the enterprise data center architecture and how we have evolved from Data Center 1.0 to Data Center 3.0. The section "Data Center 3.0 Components" covers the virtualization technologies and services that unify network, storage, compute, and virtualization platforms. The section "Data Center 3.0 Topology Components" shows how the virtualization technologies integrate with unified computing and the unified fabric.

The "Challenges in the Data Center" section describes the common server deployment challenges present in the data center. Major facility aspect issues involving rack space, power, cooling, and management are covered. Data center cabling is examined along with the data center cable considerations critical to the proper cable plant management.

Following that, the "Enterprise Data Center Infrastructure" section explores the Cisco multilayer architecture that is used for building out enterprise data centers to support blades servers, 1RU (rack unit) servers, and mainframes. Design aspects of the multilayer architecture involving data center access layer, aggregation layer, and core layer design considerations are also covered.

The chapter wraps up with several sections on virtualization. An overview of virtualization is covered along with key drivers that are pushing the adoption of virtualization in the data center. The section "Virtualization Technologies" compares the two main types of virtualization and provides several examples. Then the section "Network Virtualization Design Considerations" covers access control, path isolation, and services edge.

Enterprise DC Architectures

Over the past two decades, we have seen an evolution of data center "architectures". With Data Center 1.0, data centers were centralized, using mainframes to process and store data. The users of Data Center 1.0 used terminals to access and perform their work on the mainframes. Mainframes are still prevalent in many data centers because of the overall benefits in terms of availability, resiliency, and service level agreements (SLA).

Figure 4-1 illustrates the evolution of data center architectures from Data Center 1.0 to Data Center 3.0.

Data Center 2.0 brought client/server and distributed computing into the mainstream data center. Business applications were installed on servers and were accessed by users with client software on their PCs. Application services were distributed because of high cost of WAN links and application performance. Also, the costs of mainframes were too costly to be used as an alternative for client/server computing.

Currently, we are moving away from Data Center 2.0 and toward Data Center 3.0, where consolidation and virtualization are the key components. The cost of communication equipment is lowering, and there is an increase in computing capacities, which is driving consolidation. Data Center 3.0 centralizes the computing infrastructure and is more cost effective when compared to the distributed approach. The newer architecture takes advantage of virtualization, which results in a higher utilization of computing and network resources. In addition, the newer Data Center 3.0 architecture increases the overall return on investment (ROI) and lowers the total cost of ownership (TCO).

Data Center 3.0 Components

Figure 4-2 highlights the Cisco Data Center 3.0 components.

The architectural components of Data Center 3.0 include virtualization technologies and services that unify network, storage, compute, and virtualization platforms. These technologies and network services enable incredible flexibility, visibility, and policy enforcement, which are critical for virtualized data centers. Here are the three main components of Cisco Data Center 3.0 architecture:

• Virtualization
  • Virtual local-area network (VLAN), virtual storage-area network (VSAN), and virtual device contexts (VDC) help to segment the LAN, SAN, and network devices instances.
  • Cisco Nexus 1000V virtual switch for VMware ESX and ESXi help to deliver visibility and policy control for virtual machines (VM).
  • Flexible networking options with support for all server form factors and vendors, including support for blade servers from Cisco, Dell, IBM, and HP with integrated Ethernet and Fibre Channel switches.
• Unified fabric
  • Fibre Channel over Ethernet (FCoE) and Internet Small Computer Systems Interface (iSCSI) are two methods for implementing unified fabric in the data center over 10 Gigabit Ethernet networks.
  • FCoE is supported on VMware ESX/ESXi vSphere 4.0 and later.
  • The Cisco Catalyst, Cisco Nexus, and Cisco MDS family of switches all support iSCSI. The Cisco Nexus 5000 support unified fabric lossless operation, which improves the performance of iSCSI traffic using 10 Gigabit Ethernet.
  • The Cisco Nexus family of switches was designed to support unified fabric. Currently, the Cisco Nexus 5000 and the Nexus 4000 supports data center bridging (DCB) and FCoE. However, there are future plans for the Cisco Nexus 7000 series and the Cisco MDS family of switches to support FCoE, as well.
  • Converged network adapters (CNA) run at 10GE and support FCoE. CNAs are available from both Emulex and QLogic. Additionally, a software stack is available for certain 10GE network interfaces from Intel.
• Unified computing
  • Cisco Unified Computing System (UCS) is an innovative next-generation data center platform that converges computing, network, storage, and virtualization together into one system.
  • Integrates lossless 10GE unified network fabric with x86 architecture-based servers.
  • Allows for Cisco Virtual Interface Card to virtualize your network interfaces on your server.
  • Offers Cisco VN-Link virtualization.
  • Supports Extended Memory Technology patented by Cisco.
  • Increases productivity with just-in-time provisioning using service profiles.

Data Center 3.0 Topology Components

Figure 4-3 shows the Cisco Data Center 3.0 topology.

At the top layer, we have virtual machines which are software entities that have hardware level abstraction capable of running a guest OS on top of a resource scheduler also known as a hypervisor.

Within the unified computing resources, the service profile defines the identity of the server. The identity contains many items such as memory, CPU, network cards, storage information, and boot image.

10 Gigabit Ethernet, FCoE, and Fibre Channel technologies provide the unified fabric and is supported on the Cisco Nexus 5000. FCoE is one of the key technologies that allow native Fibre Channel frames to be used on 10G Ethernet networks.

Virtualization technologies such as VLANs and VSANs provide for virtualized LAN and SAN connectivity by logically segmenting multiple LANs and SANs on the same physical equipment. Each VLAN and VSAN operates independently from one another.

At the lowest layer, we have virtualized hardware where storage devices can be virtualized into storage pools and network devices are virtualized using virtual device contexts (VDC).

Challenges in the DC

In the data center, server deployments are of great concern along with facilities and network equipment. Here are some of the challenges that must be dealt with when deploying servers:

• Power required
• Physical rack space usage
• Limits to scale
• Management (resources, firmware)
• Server security
• Virtualization support
• Management effort required

Server growth is consistently rising which is requiring more power, which is driving the need for energy efficiency for most data center server deployments. Although rack servers are low cost and provide high performance, unfortunately they take up space and consume a lot of energy to operate. Because both rack space and power cost money, efficiency gains need to be considered in these areas.

Blade servers provide similar computing power when compared to rack mount servers, but require less space, power, and cabling. The chassis in most blade servers allows for shared power, Ethernet LAN, and Fibre Channel SAN connections, which reduce the number of cables needed.

With both rack-mounted servers and blade servers, server virtualization software provides for better utilization of hardware resources, which requires less physical hardware to deploy servers, which in turn increases efficiency. Server virtualization also enables server scalability because more rack and cabinet space is available to deploy new ESX hosts running additional virtual machines.

Server management is a key element for deploying servers, and there are solutions available from OEMs such as Integrated Lights Out (ILO) and VMware Infrastructure Client. These products ease the management of larger server deployments and provide for secure remote management capabilities.

Data Center Facility Aspects

Multiple facility considerations go into the design and planning for a new data center build out.

During the planning sessions, data center architectural and mechanical specifications help define the following:

• How much space will be available
• How much load the floor can support
• The power and cooling capacity that will be available
• The cabling plant that will be needed and how to manage it

The facility also needs to meet certain environmental conditions, and the data center equipment selections process dictates the operating temperatures and humidity levels that need to be maintained in the data center.

Another important consideration is physical security. Because the data center usually stores data that needs to be secured from third parties, access to the site needs to be well controlled. In addition, fire suppression and alarm systems should be in place to protect equipment and data from natural disasters and theft.

Because the data center facilities are limited in capacity, they need to be designed properly to allow for the best use of employee space for today and into the future.

Most companies must now adhere to regulatory compliance, including environmental requirements, and provide disaster recovery in some form to enable business continuity. Data centers need to provide an infrastructure that can recover network communications, data, and applications and provide high availability.

To build a reliable data center that maximizes the investment, the design needs to be considered early in the building development process. It is important to include team members in several area of expertise, including telecommunications, power, architectural, and heating, ventilating, and air conditioning (HVAC). Each team member needs to work together to ensure that the designed systems interoperate most effectively. The design of the data center needs to incorporate current requirements and support future growth.

Careful planning and close attention to design guidelines is crucial for the data center build out to be successful. Missing critical aspects of the design can cause the data center to be vulnerable to early obsolescence, which can impact data center availability and lead to a loss of revenue or increased cost to remediate.

Table 4-2 describes a number of data center facility considerations.

Table 4-2. Summary of Data Center Facility Considerations

Data Center Facility Considerations	Description
Architectural and mechanical specifications	Space available Load capacity Power and cooling capacity Cabling infrastructure
Environmental conditions	Operating temperature Humidity level
Physical security	Access to the site Fire suppression Security Alarms
Capacity limits	Space for employees
Compliance and regulation	Payment Card Industry (PCI), Sarbannes-Oxley (SOX), and Health Insurance Portability and Accountability Act (HIPAA)

Data Center Space

The space that the data center occupies makes up the physical footprint and helps answer many questions, including how to size the overall data center, where to position servers, how to make it flexible for future growth, and how to protect the valuable equipment inside.

The data center space element defines the number of racks for servers and telecommunications equipment that can be installed. The floor loading is affected by the rack weight after the racks are populated with equipment. Careful planning is needed to ensure that the floor loading is sufficient for current and future needs of the data center.

Selecting the proper size of the data center has a great influence on the cost, longevity, and flexibility of the data center. Although estimating the size of the data center is challenging, it is also critically importance that it be done correctly.

Several factors need to be considered, including the following:

• The number of employees who will be supporting the data center
• The number of servers and the amount of storage gear and networking equipment that will be needed
• The space needed for non-infrastructure areas:
  • Shipping and receiving
  • Server and network staging
  • Storage rooms, break rooms, and bath rooms
  • Employee office space

Keep in mind that if the data center is undersized it will not sufficiently satisfy compute, storage, and network requirements and will negatively impact productivity and cause additional costs for expansion. On the flip side, a data center that is too spacious is a waste of capital and recurring operational expenses.

Right-size data center facilities consider the placement of infrastructure and equipment; and if properly planned, the data center can grow and support the organization into the future without costly upgrades or relocations.

Here are some other rack and cabinet space considerations to keep in mind:

• Weight of the rack and equipment
• Heat expelled from equipment
• Amount and type of power needed
  • Automatic transfer switch for equipment that has single power supplies
  • Uninterruptible power supplies (UPS)
  • Redundant power distribution units (PDU)
• Loading, which determines what and how many devices can be installed

Data Center Power

The power in the data center facility is used to power cooling devices, servers, storage equipment, the network, and some lighting equipment. Cooling down the data center requires the most power, next to servers and storage.

Because many variables make up actual power usage, determining power requirements for equipment in the data center can prove difficult. In server environments, the power usage depends on the computing load place on the server. For example, if the server needs to work harder by processing more data, it has to draw more AC power from the power supply, which in turn creates more heat that needs to be cooled down.

The desired reliability drives the power requirements, which may include multiple power feeds from the power utility, UPS, redundant power circuits, and diesel generators. Depending on the options chosen, various levels of power redundancy can affect both capital and recurring operating expenses. Determining the right amount of power redundancy to meet the requirements takes careful planning to ensure success.

Estimating the power capacity needed involves collecting the requirements for all the current equipment, including the future requirements of the equipment for the data center. The complete power requirements must encompass the UPS, generators, HVAC, lighting, and all the network, server, and storage equipment.

Figure 4-4 shows an example of data center power usage.

The designed power system should include electrical components such as PDUs, circuit breaker panels, electrical conduits, and wiring necessary to support the desired amount of physical redundancy. The power system also needs to provide protection for utility power failures, power surges, and other electrical problems by addressing the power redundancy requirements in the design.

Here are some key points related to data center power:

• Defines the overall power capacity.
• Provides physical electrical infrastructure and addresses redundancy.
• Power is consumed by the following:
  • Cooling
  • Servers
  • Storage
  • Network
  • Conversion and lighting

Data Center Cooling

Devices in the data center produce variable amounts of heat depending on the device load. Heat overtime decreases the reliability of the data center devices. Cooling is used to control the temperature and humidity of the devices, and it is applied to zones, racks, or individual devices.

Environmental conditions need to be considered and measured by using probes to measure temperature changes, hot spots, and relative humidity.

A major issue with high-density computing is overheating. There are more hot spots, and therefore more heat overall is produced. The increase in heat and humidity threatens equipment life spans. Computing power and memory requirements demand more power and thus generate more heat output. Space-saving servers increase the server density possible in a rack, but keep in mind that density = heat. It might not be a big deal for one chassis at 3 kilowatt (kW), but with five or six servers per rack, the heat output increases to 20 kW. In addition, humidity levels can affect static electricity in the data center. So, it is recommended that relative humidity level be in the range of 40 percent to 55 percent. High levels of static electricity can cause damage to data center equipment.

Proper airflow is required to reduce the amount of heat generated by the high-density equipment. Sufficient cooling equipment must be available to produce acceptable temperatures within the data center. The cabinets and racks should be arranged in the data center with an alternating pattern of "cold" and "hot" aisles. The cold aisle should have equipment arranged face to face, and the hot aisle should have equipment arranged back to back. In the cold aisle, there should be perforated floor tiles drawing cold air from the floor into the face of the equipment. This cold air passes through the equipment and flushes out the back into the hot aisle. The hot aisle does not have any perforated tiles, and this design prevents the hot air from mixing with the cold air.

Figure 4-5 illustrates the alternating pattern of cold and hot aisles along with airflow.

For equipment that does not exhaust heat to the rear, here are some other cooling techniques:

• Block unnecessary air escapes to increase airflow.
• Increase the height of the raised floor.
• Spread out equipment into unused racks.
• Use open racks rather than cabinets where security is not a concern.
• Use cabinets with mesh fronts and backs.
• Custom perforated tiles with larger openings.

Data Center Heat

Blade server deployments allow for more efficient use of space for servers, which is good, but there is also an increased amount of heat per server, which requires more cooling to maintain consistent temperatures.

The data center design must address the increased use of high density servers and the heat that they produce. During the data center design, considerations for cooling need to be taken into account for the proper sizing of the servers and the anticipated growth of the servers along with their corresponding heat output.

Here are some cooling solutions to address the increasing heat production:

• Increase the number of HVAC units.
• Increase the airflow through the devices.
• Increase the space between the racks and rows.
• Use alternative cooling technologies, such as water-cooled racks.

Data Center Cabling

The cabling in the data center is known as the passive infrastructure. Data center teams rely on a structured and well-organized cabling plant. Although the active electronics are crucial for keeping server, storage, and network devices up and running, the physical cabling infrastructure is what ties everything together. The cabling in the data center terminates connections between devices and governs how each device communicates with one another.

Cabling has several key characteristics, such as the physical connector, media type, and cable length. Copper and fiber-optic cables are commonly used today. Fiber-optic cabling allows for longer distances and is less prone to interference than copper cabling. The two main types of optical fiber are single-mode and multi-mode. Copper cabling is widely available, costs less, and generally covers shorter distances (up to 100 meters, about 328 feet). Typical copper cabling found in the data center is CAT 5e/CAT 6 with RJ-45 connectors.

Keep in mind that the emerging 10GBASE-T standard requires CAT6A twisted-pair cabling to support distances up to 100 meters.

It is important for cabling to be easy to maintain, abundant and capable of supporting various media types and requirements for proper data center operations.

Cable management and simplicity is affected by the following:

• Media selection
• Number of connections
• Type of cable termination organizers
• Space for cables on horizontal and vertical cable trays

These considerations must to be addressed during the data center facility design (for the server, storage, network, and all the associated technologies that are going to be implemented).

Figure 4-6 shows an example of cabling that is out of control.

Figure 4-7 shows the proper way to manage copper cabling.

The cabling infrastructure needs to avoid the following pitfalls:

• Inadequate cooling due to restricted airflow
• Outages due to accidental disconnect
• Unplanned dependencies resulting in more downtime
• Difficult troubleshooting options

For example, using under-floor cabling techniques, especially with a high number of power and data cables can restrict proper airflow. Another disadvantage with this approach is that cable changes require you to lift floor tiles, which changes the airflow and creates cooling inefficiencies.

One solution is a cable management system above the rack for server connectivity. Cables should be located in the front or rear of the rack to simplify cable connections. In most service provider environments, cabling is located in the front of the rack.

Enterprise DC Infrastructure

Today's enterprise data center design follows the Cisco multilayer architecture, which includes DC core, DC aggregation, and DC access layers. This multitier model is the most common model used in the enterprise and it supports blade servers, single rack unit (1RU) servers, and mainframes.

Figure 4-8 provides a high-level overview of an enterprise data center infrastructure.

At the edge of the data center infrastructure is the access layer. The data center access layer needs to provide physical port density and both Layer 2 and Layer 3 services for flexible server connectivity options.

The data center aggregation layer ties the DC core and DC access layers together, which provides hierarchy for security and server farm services. Security services such as access control lists (ACL), firewalls, and intrusion prevention systems (IPS) should be implemented in the data center aggregation layer. In addition, server farm services such as content switching, caching, and Secure Sockets Layer (SSL) offloading should be deployed in the data center aggregation. Both the data center aggregation and core layers are commonly implemented in pairs for redundancy, to avoid single points of failure.

Defining the DC Access Layer

The data center access layer's main purpose is to provide Layer 2 and Layer 3 physical port density for various servers in the data center. In addition, data center access layer switches provide high-performance, low-latency switching and can support a mix of oversubscription requirements. Both Layer 2 and Layer 3 access (also called routed access) designs are available, but most data center access layers are built using Layer 2 connectivity. The Layer 2 access design uses VLAN trunks upstream, which allows data center aggregation services to be shared across the same VLAN and across multiple switches. Other advantages of Layer 2 access are support for NIC teaming and server clustering that requires network connections to be Layer 2 adjacent or on the same VLAN with one another.

Figure 4-9 highlights the data center access layer in the overall enterprise architecture.

The Spanning Tree Protocol (STP) manages physical loops that are present in the Layer 2 design. Currently, the recommended STP mode is Rapid per-VLAN Spanning Tree Plus (RPVST+), which ensures a logical loop-free topology and fast convergence.

New routed access designs aim to contain Layer 2 locally to avoid the use of the STP. With routed access designs, the default gateway function needs to be provided because the access switch becomes the first-hop router in the network.

Designs with both Layer 2 and Layer 3 access provide flexibility for multiple server solutions to be supported, including 1RU servers and modular blade server chassis.

Here are some of the data center access layer benefits:

• Provides port density for server farms
• Supports single homed and dual homed servers
• Provides high-performance, low-latency Layer 2 switching
• Supports mix of oversubscription requirements

Defining the DC Aggregation Layer

The data center aggregation (distribution) layer aggregates Layer 2/Layer 3 links from the access layer and connects using upstream links to the data center core. Layer 3 connectivity is typically implemented between the data center aggregation and the data center core layers. The aggregation layer is a critical point for security and application services. The Layer 4 security and application services in the data center aggregation layer include server load balancing, SSL offloading, firewalling, and IPS services. These services maintain connection and session state for redundancy purposes and are commonly deployed in pairs using Cisco Catalyst 6500 service modules. This design reduces the total cost of ownership (TCO) and eases the management overhead by simplifying the number of devices that must be managed.

The highlighted section in Figure 4-10 illustrates the data center aggregation layer.

Depending on the requirements of the design, the boundary between Layer 2 and Layer 3 can be in the multilayer switches, firewalls, or content switching devices in the aggregation layer. Multiple aggregation layers can be built out to support separate network environments, such as production, test, and PCI infrastructure, each with its own security zone and application services. First-hop redundancy protocols Hot Standby Router Protocol (HRSP) and Gateway Load Balancing Protocol (GLBP) are commonly used in the aggregation layer. Many aggregation designs include positioning STP primary and secondary root bridges to help control the loop-free topology and support a larger STP processing load.

Here are some of the data center aggregation layer benefits:

• Aggregates traffic from DC access and connects to DC core.
• Supports advanced application and security services.
• Layer 4 services include firewall, server load balancing, SSL offload, and IPS.
• Large STP processing load.
• Highly flexible and scalable.

Defining the DC Core Layer

The data center core connects the campus core to the data center aggregation layer using high-speed Layer 3 links. The core is a centralized Layer 3 routing layer in which one or more data center aggregation layers connect. The data center networks are summarized, and the core injects the default route into data center aggregation. The data center core also needs to support IP multicast to provide connectivity to the growing use of IP multicast applications.

The data center core layer is a best practice component of larger data center networks. Smaller data centers may use a collapsed core design combining the aggregation layer and core layers together. However, if you are building a greenfield data center, it is recommended to implement a data center core in the beginning to avoid network downtime later. Table 4-3 shows some drivers to help you decide whether a data center core is appropriate for your design.

Table 4-3. Data Center Core Drivers

Data Center Core Drivers	Description
10 Gigabit Ethernet density	Are there enough 10GE ports to connect campus core to multiple data center aggregation layers?
Administrative domains and policies	Separate cores help to isolate campus distribution from DC aggregation for troubleshooting and quality of service/access control list (QoS/ACL) policies.
Future growth	The impact and downtime from implementing a core at a later date make it worthwhile to install sufficent core layers in the beginning.

The highlighted section in Figure 4-11 illustrates the data center core layer.

Here are some of the data center core characteristics:

• Low-latency switching
• Distributed forwarding architecture
• 10 Gigabit Ethernet
• Scalable IP multicast support

Virtualization Overview

As the demand for IT to do more with less while increasing efficiency has risen, virtualization has become a critical component in most enterprise networks. Virtualization technologies allow a physical device to share its resources by acting as multiple versions of itself. Other forms of virtualization can enable multiple physical devices to logically appear as one.

Virtualization is a critical component of the Cisco network architectures for the enterprise data center and is changing the way data centers are architected. The use of virtualization improves network efficiency, provides enhanced flexibility, and reduces operational expenses.

Challenges

Network designers face many challenges that are driving the need to deploy virtualization technologies in the network. Data centers are growing rapidly, and these challenges directly impact the profitability of the business.

Take a look at some of the key driving forces for virtualization adoption in Table 4-4.

Table 4-4. Virtualization Key Drivers

Virtualization Driving Forces	Description
Operational cost	Need to reduce rising cost of powering and cooling devices in the DC while getting more productivity
Reduce the number of physical devices	DC consolidation of assets performing individual tasks
Traffic isolation	Logical, separate user groups secured from other groups on the same network
Increased performance/price ratio	Eliminate underutilized hardware that exhibits poor performance/price ratio

Defining Virtualization and Benefits

Virtualization is an umbrella term used to represent several different technologies. Virtualization technologies share a common theme in their ability to abstract logical elements from hardware (applications or operating systems) or networks (LANs and SANs) and run them in a virtual state. Virtualization brings many benefits, from consolidation to increased efficiency.

Here are some of the common benefits achieved through virtualization techniques:

• Better use of computing resources, higher server densities, and simplified server migrations
• Provides flexibility for ease of management for adds, reassignments, or repurposing resources
• Separation of users groups on the same physical network, enabling traffic isolation
• Ability to provide per-department security policies
• Reduction in power and space required
• Increased uptime and reduced operational costs

Types of Virtualization

Enterprise networks consist of two main types of virtualization technologies groupings, called network virtualization and device virtualization:

• Network virtualization encompasses logical isolated network segments that share the same physical infrastructure. Each segment operates independently and is logically separate from the other segments. Each network segment appears with its own privacy, security, independent set of policies, QoS levels, and independent routing paths.

  Here are some examples of network virtualization technologies:

  • VLAN: Virtual local-area network
  • VSAN: Virtual storage-area network
  • VRF: Virtual routing and forwarding
  • VPN: Virtual private network
  • vPC: Virtual Port Channel

• Device virtualization allows for a single physical device to act like multiple copies of itself. Device virtualization enables many logical devices to run independently of each other on the same physical piece of hardware. The software creates virtual hardware that can function just like the physical network device. Another form of device virtualization entails using multiple physical devices to act as one logical unit.

  Here are some examples of device virtualization technologies:

  • Server virtualization: Virtual machines (VM)
  • Cisco Application Control Engine (ACE) context
  • Virtual Switching System (VSS)
  • Cisco Adaptive Security Appliance (ASA) firewall context
  • Virtual device contexts (VDC)

Virtualization Technologies

Virtualization is built from abstracting logical entities from pooled physical resources. The Cisco network architectures for the enterprise data center contains many forms of network and device virtualization technologies.

Figure 4-12 illustrates the many virtualization technologies in use today.

VSS

Virtual Switching System (VSS) is a network virtualization technology that allows two physical Cisco Catalyst 6500 series switches to act as a single logical virtual switch. The VSS increases operational efficiencies and scales bandwidth up to 1.4 Tb/s. This technology is very similar to StackWise technology used with the Cisco Catalyst 3750 series product line, which enables switches stacked together to operate as one and use a single command-line interface (CLI) for management. However, VSS is limited to two physical chassis connected together.

VRF

Virtual routing and forwarding (VRF) is a routing virtualization technology that creates multiple logical Layer 3 routing and forwarding instances (route tables) that can function on the same physical router. In Multiprotocol Label Switching (MPLS) VPN environments, the use of VRF technology plays a major role by allowing multiple networks to coexist on the same MPLS network. The routing information is contained inside the VRF and is visible only to routers participating in the same VRF. Because the routing information with VRF is separated, duplicate IP addressing schemes can be used.

vPC

Virtual Port Channel (vPC) technology works by combining two Cisco Nexus 7000 series switches or two Cisco Nexus 5000 series switches with 10GE links, which are then represented to other switches as a single logical switch for port channeling purposes. With vPC, the spanning-tree topology appears loop-free, although multiple redundant paths are present in the physical topology.

Device Contexts

Device contexts enable a single physical network device to host multiple virtual network devices. Each device context is an independent configuration with its own policy, network interfaces, and management accounts. The virtualized contexts that run on a single network device operate similarly to standalone network devices. Most of the same features present on the physical device are also supported on the individual device contexts.

The following Cisco network devices support the use of device contexts:

• Cisco Nexus 7000 series switches (VDC)
• Cisco Adaptive Security Appliance (ASA) firewall
• Cisco Catalyst 6500 Firewall Services Module (FWSM)
• Cisco Application Control Engine Appliance
• Cisco Catalyst 6500 Application Control Engine Module
• Cisco Intrusion Prevention System (IPS)

Server Virtualization

The use of server virtualization has exploded onto the market over the past several years and can be found in most data center environments. Server virtualization is a software technique that abstracts server resources form the hardware to provide flexibility and to optimize the usage of the underlying hardware. As a result, many data center applications are no longer bound to bare-metal hardware resources.

The server virtualized hypervisor provides the foundation for the virtualized environment on the host. The hypervisor controls the hardware and physical resources that can be allocated to virtual machines running on the host. This makes the VMs unaware of the physical hardware, but they can use CPUs, memory, and network infrastructure as shared pools available through the virtualization process.

The following represents several server virtualization vendors and their associated products:

• VMware ESX Server
• Citrix XenServer
• Microsoft Hyper-V

Network Virtualization Design Considerations

Network solutions are needed to solve the challenges of sharing network resources but keeping users totally separate from one another. Although the users are separate, we still need to ensure that the network is highly available, secure, and can scale along with the business growth. Network virtualization offers solutions to these challenges and provides design considerations around access control, path isolation, and services edge.

Access Control

Access needs to be controlled to ensure that users and devices are identified and authorized for entry to their assigned network segment. Security at the access layer is critical for protecting the network from threats, both internal and external.

Path Isolation

Path isolation involves the creation of independent logical network paths over a shared network infrastructure. MPLS VPN is an example of path-isolation technique where devices are mapped to a VRF to access the correct set of network resources. Other segmentation options include VLANs and VSANs, which logically separate LANs and SANs. The main goal when segmenting the network is to improve the scalability, resiliency, and security services as with non-segmented networks.

Services Edge

The services edge refers to making network services available to the intended users, groups, and devices with an enforced centralized managed policy. Separate groups or devices occasionally need to share information that may be on different VLANs, each with corresponding group policies. For example, traffic from the sales VLAN might need to talk to the engineering VLAN, but it needs to go through the firewall to permit the traffic and might even be tied to certain hours of the day. In such cases, the network should have a central way to manage the policy and control access to the resources. An effective way to address policy enforcement is to use an FWSM in a Cisco Catalyst 6500 series switch providing firewall services for the data center.

Table 4-5 describes network virtualization considerations.

Table 4-5. Network Virtualization Design Considerations

Network Virtualization Consideration	Description
Access control	Ensures users and devices are recognized, classified, and authorized for entry to their assigned network segments
Path isolation	Provides independent logical traffic paths over shared network
Services edge	Ensures the right services are accessible the intended users, groups, or devices

References and Recommended Readings

Module 3 (Designing Basic Campus and Data Center Networks)—Designing for Cisco Internetwork Solution Course (DESGN) 2.1.

Cisco Design Zone for Data Centers, www.cisco.com/en/US/netsol/ns743/networking_solutions_program_home.html.

Data Center Design – IP Network Infrastructure, www.cisco.com/en/US/docs/solutions/Enterprise/Data_Center/DC_3_0/DC-3_0_IPInfra.html#wp1037111.

Cisco Data Center Infrastructure 2.5 Design Guide, www.cisco.com/en/US/docs/solutions/Enterprise/Data_Center/DC_Infra2_5/DCInfra_2.html.

Security and Virtualization in the Data Center, www.cisco.com/en/US/docs/solutions/Enterprise/Data_Center/DC_3_0/dc_sec_design.html#wp1056029.

Examples of Good Cable Management, www.geekshout.com/media/photos/9-examples-of-good-cable-management-and-1-bad-one/.