Analyzing server hardware

1
(Skill 1)
Analyzing Server and Workstation Hardware

• Analyzing server hardware
• Check the minimum and recommended requirements of
  the target operating system (Table 1-1)
• Future planning
• Make sure servers can meet any additional demands
  you may design into the infrastructure
• Make sure servers will scale to meet future needs
  based on anticipated growth during the targeted
  server life cycle

2
(Skill 1)
Analyzing Server and Workstation Hardware (2)

• Analyzing client workstations
• Check the minimum and recommended requirements of
  the target operating system (Table 1-2)
• Make sure client computers can support necessary
  operating system upgrades

3
(Skill 1)
Analyzing Server and Workstation Hardware (3)

• Analyze load under which computers will be placed
• Analyze performance logs over a one- to two-week
  interval (baseline)
• Note key bottlenecks
• Examine the current level of service
• Interview key personnel to determine if level of
  service is adequate

4
(Skill 1)
Analyzing Server and Workstation Hardware (4)

• Performing a pilot upgrade
• Use the hardware and software you are planning to
  recommend
• Track the performance levels over a reasonable
  period of time
• Make sure the recommended hardware specifications
  meet the desired performance level

5
(Skill 1)
Figure 1-2 Calculating processor capacity
6
(Skill 2)
Examining Software Requirements

• Verify the operating system is compatible with
  existing software
• If business-critical or custom software is in
  use, you may need to contact outside vendors or
  programmers
• Make sure the minimum hardware requirements for
  software packages are met
• Verify all software licensing requirements are
  met
• When designing an end-to-end solution, licensing
  requirements can significantly impact projects
  budget

7
(Skill 2)
Figure 1-3 Steps for examining software
requirements
8
(Skill 3)
Examining Current Network Hardware

• Make sure network hardware meets the demands of
  the new design
• Standard types of hardware
• Hubs
• Multistation access units (MAUs)
• Wireless access points (WAPs)
• Layer 2 switches and bridges
• Routers and Layer 3 switches

9
(Skill 3)
Examining Current Network Hardware (2)

• Hubs (also known as multiport repeaters)
• Typically used in Ethernet networks
• Provide only one collision domain
• Have been rendered nearly obsolete by Layer 2
  switches

10
(Skill 3)
Examining Current Network Hardware (3)

• Multistation access units (MAUs)
• Connection devices for Token Ring networks
• Function is very similar to that of a hub

11
(Skill 3)
Examining Current Network Hardware (4)

• Wireless access points (WAPs)
• Essentially hubs for wireless LANs (WLANs)
• Capabilities
• Provide a connection point to a wired network
• Expand the range of a wireless network
• In some instances, provide security

12
(Skill 3)
Examining Current Network Hardware (5)

• Layer 2 switches and bridges
• Filter packets based on the media access control
  (MAC) address and forward them out of the
  appropriate ports
• Main difference between a Layer 2 switch and a
  bridge is in the number of ports
• Bridges typically only have two ports
• Switches have more than two
• Typically provide a collision domain for each port

13
(Skill 3)
Examining Current Network Hardware (6)

• Layer 3 switches and routers
• Divide the network into broadcast domains
• Primary difference between Layer 3 switch and
  router is simply in classification

14
(Skill 3)
Examining Current Network Hardware (8)

• Performance factors
• For hubs and MAUs, the primary concern is
  available bandwidth
• For full duplex networks, make sure the bandwidth
  will not be exceeded in the new design
• For half-duplex networks, make sure no more than
  40 to 50 of total the bandwidth will be exceeded
  in the new design

15
(Skill 3)
Examining Current Network Hardware (9)

• Performance factors
• Estimating bandwidth requirements
• Examine current bandwidth usage
• Determine the amount of headroom left
• Examine services and features to be included in
  the new design
• Determine how much additional bandwidth each
  requires
• In most networks, there will be considerable load
  placed upon the network during peak periods

16
(Skill 3)
Examining Current Network Hardware (12)

• Performance factors for Layer 2 switches and
  bridging devices
• Packets per second (PPS) limit
• Typically the best indicator of a switchs
  potential performance
• Defines the maximum rate the entire switching
  fabric can support
• Low-end switches typically have a PPS rating of
  100,000 to 1 million
• High-end switches typically have a PPS rating of
  10 million or more

17
(Skill 3)
Examining Current Network Hardware (13)

• Performance factors for Layer 2 switches and
  bridging devices
• Media access control (MAC) address limit
• If a switch is not capable of supporting the
  number of MAC addresses available in the
  broadcast domain, flooding occurs
• Low-end switches typically have a MAC address
  limit of 512 to 2,000 addresses
• High-end switches typically have a MAC address
  limit of 16,000 or more addresses

18
(Skill 3)
Examining Current Network Hardware (14)

• Performance factors for Layer 2 switches and
  bridging devices
• General considerations
• Make sure the switch can adequately handle
  traffic being sent to it during peak periods
• Use a third-party Simple Network Management
  Protocol (SNMP) suite if possible
• Estimate a switchs potential maximum performance
  by multiplying the PPS by the average packet size
  (in bits) on the network

19
(Skill 3)
Examining Current Network Hardware (15)

• Performance factors for Layer 3 devices
• Performance analysis is similar to the analysis
  for Layer 2 switches
• Make sure the processor and RAM are adequate to
  meet needs of design
• An increase in the number of packets that must be
  routed increases processor load
• An increase in the number of subnets in the
  routers routing table increases RAM requirements

20
(Skill 3)
Figure 1-4 Network hardware
21
(Skill 3)
Figure 1-6 Standard recommendations for
increasing network security
22
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (2)

• Transmission Control Protocol/Internet Protocol
  (TCP/IP)
• Most common suite in use
• Robust, scalable, routable, and
  vendor-independent
• Active Directory relies on TCP/IP, which means in
  most cases, you will be designing for TCP/IP
  networks

23
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (3)

• Internetwork Packet Exchange/Sequenced Packet
  Exchange (IPX/SPX)
• Used with Novell NetWare line of network
  operating systems
• Robust, scalable, and routable
• NetWare 5 and later is capable of using TCP/IP
  for network communications, but older versions
  require IPX/SPX support

24
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (4)

• Internetwork Packet Exchange/Sequenced Packet
  Exchange (IPX/SPX)
• If integrating or migrating a NetWare network to
  Active Directory
• Convert the entire network to TCP/IP if possible
• If not possible, operate the two protocol suites
  on a single network or use a gateway service such
  as Gateway Services for NetWare (GSNW) to provide
  access to legacy Novell systems

25
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (5)

• NetBIOS Extended User Interface (NetBEUI)
• Most commonly associated with older Microsoft
  operating systems
• Rarely used in large, modern networks
• Not robust, scalable, or routable
• Migrate these networks to TCP/IP
• If not possible, you can run both TCP/IP and
  NetBEUI simultaneously, but it will impact
  network traffic considerably

26
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (8)

• Routing topologies
• Star
• Ring
• Hybrid mesh
• Full mesh

27
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (9)

• Star topology
• Also known as hub-and-spoke
• Most common type
• Advantages reduction in complexity and cost
• Disadvantages emphasizes a single point of
  failure and lacks redundancy

28
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (10)

• Ring topology
• Not particularly common
• Advantages no single point of failure, not
  particularly complex
• Disadvantage additional number of links present
  adds additional cost

29
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (11)

• Hybrid mesh topology
• Common for large organizations and Internet
  service providers (ISPs)
• Does not fit all requirements of the full mesh
  topology, but provides additional redundancy
• Costs scale according to level of redundancy
  chosen

30
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (12)

• Full mesh topology
• Also known as true mesh
• Each router has an individual connection to every
  other router
• Number of links required can be very high
• Rare due to the cost and complexity involved in
  implementation

31
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (13)

• Routing protocols
• Designed to automatically locate and utilize
  secondary paths when a primary path fails
• Only required if a network has redundant paths
  if it does not, use static routing
• Routing protocols provided in Windows Server 2003
  and Windows 2000 Server
• Routing Information Protocol (RIP) versions 1 and
  2
• Open Shortest Path First (OSPF)

32
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (14)

• Routing Information Protocol (RIP) version 1
• Advertises the entire routing table at periodic
  intervals (30 seconds by default) to all
  configured interfaces
• Inefficient from a bandwidth perspective
• Does not provide high level of scalability
• Hop count (only metric used) limited to 15 hops
• Does not support
• Password protection of routing updates
• Variable Length Subnet Mask (VLSM)
• Classless Interdomain Routing (CIDR)

33
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (15)

• Routing Information Protocol (RIP) version 2
• Improved version of RIP version 1
• Added support
• Password protection of routing updates
• Multicasting of updates
• Variable Length Subnet Mask (VLSM)
• Classless Interdomain Routing (CIDR)
• Retains inefficient update mechanism and limited
  scalability

34
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (16)

• Open Shortest Path First (OSPF)
• Highly bandwidth efficient and scalable
• Advertises only when initially booted and when
  the state of a link changes
• Creates map of network, enabling it to easily and
  efficiently chose best path
• Uses link speed as its metric, which makes path
  selection more accurate
• Almost always recommended over RIP version 1 or 2

35
(Skill 4)
Examining the Current Protocol Requirements and
Routing Infrastructure (17)

• OSPF supports
• Password protection of routing updates
• Multicasting of updates
• Variable Length Subnet Mask (VLSM)
• Classless Interdomain Routing (CIDR)
• OSPF limitations
• More complex than RIP and uses more processor and
  RAM resources on router
• Cannot be used over demand-dial routing
  connections

36
(Skill 4)
Figure 1-8 Using a gateway to connect to NetWare
systems