Howard Goldstein Associates, Inc.

Business, Technology & Education.

Course Title:  Storage Networking Design, Performance, Planning, & Troubleshooting

 

Vendor: Howard Goldstein Associates, Inc.        

Email: Howard.Goldstein@HGAI.Com

 

Locations: Private classes available.                   Phone: 303-554-0755

 

Duration: 2 and 4-day versions of the course

 

Type: Lecture/Demo       Course URL: www.hgai.com

 

Course Description / Overview / Expected Outcome:

 

Now that Storage Networking is popularized with a myriad of topologies and technologies the time has come to discuss a hype-free, technical perspective involved in Planning, Design, Performance and Troubleshooting implementations.

 

Examine design alternatives exposing strengths and weaknesses of these approaches.  This course looks at ways to scale designs and maximize infrastructure investments.  “Does it really make sense to connect a low “offered load” web server to a SAN Director or switch when the switch port maximum utilization is only 20% and the cost per port is $10K -$20K?”  That’s asset waste, a direct conflict to the SAN total asset utilization benefit.  “Does it really pay to rob Peter and pay Paul?”  “Is ROI really a given or is this more hype?”  There are alternatives to the over-engineered approaches based on inappropriate Storage Networking assumptions.  “Does the “Any-to-Any” design implemented in Internet and Intranets work efficiently for Storage Networking?”

 

We examine the technical performance drivers and tuning opportunities in end-to-end approaches including Storage Network, server and storage controller components giving a clear, analytical, quantitative approach from the hour to nanosecond level.  “What is can I realistically achieve with Storage Networking regardless of the latest port bandwidth improvement offered?” 

 

Finally, examine Storage Networking problems and ways to solve them. “Where has my storage LUN gone?” “Why is my file corrupted?” “Why constant failover once I installed my SAN?”  Examine and demonstrate use of tools like protocol analyzers, problem detection and performance analysis functions.

 

Course Objectives:

 

Identify what Storage Networking is and how it relates to storage

 

Discuss a hype-free, technical and business perspective in Planning, Design, Performance and Troubleshooting Storage Networks

 

Examine design alternatives exposing strengths and weaknesses of Storage Network Designs

 

Examine the technical performance drivers and tuning opportunities in end-to-end approaches including Storage Network, Server, Storage Controller components giving a clear, analytical, quantitative approach from the hour to nanosecond level

 

Maximize port and bandwidth utilization

 

Examine a top-down approach to information flow requirements levied on a storage network and how various storage network technologies meet those requirements.

 

Explore the many topologies and the underlying technologies and protocols that can be used in storage networks including Point-to-Point, Arbitrated Loop and Switched Fabric and the trade offs in cost/benefit where these topologies can be applied today.

 

Examine the current and future trends in Storage Networking technologies and protocols and the developments that are occurring to improve the benefits that can be achieved.

 

Use a storage networking top-down planning & design process to select and evaluate storage network topologies, technologies and products

 

Gather information flow requirements in current and future implementation environments

 

Examine and use state-of-the art tools that address storage network planning, design, performance and troubleshooting

 

Understand the value in separating the planning, design, performance and troubleshooting of the physical transport network and the application solutions, servers and storage systems that levy requirements on the storage network.

 

Analyze high and low performance in DAS, NAS and SAN environments and to understand how to tune and optimize performance in a storage network

 

Learn and use troubleshooting techniques for a storage network implementation

 

 

What’s exciting about it, why it’s important, where it’s going: 

 

· It provides a one-of-a-kind workshop with an industry expert who has real-world experience.

· It provides a vendor-neutral view of storage network design with a focus leveraging your investment in network infrastructure, existing and future storage and server assets.

· It identifies the true impacts to performance and provides useful insights into ways of achieving a high performance storage network.

· It examines tools, tips and tricks discovered over the last 5 years while working in storage networking.

· Understand the kinds of problems that can occur in a storage network and ways of solving these problems quickly and efficiently.

· It puts you in position to understand the technology issues for storage networks and to create implement and manage a better storage network for your environment.

 

Who Should Attend:

 

· This course is targeted towards anyone who designs, implements, manages, specifies or selects Storage Networking technologies.

· This includes IS/IT technical staff and managers, product developers, systems integrators, systems engineers and technical marketing personnel.

· Anyone who wants to understand more about storage networking solutions and technologies

· This course is designed to provide technical and management IT staff a short but comprehensive understanding of Storage Networking.

· Product and project teams that are involved with applications, systems, storage and end users will benefit from this course.

· Developers, integrators, engineers, administrators, managers, marketing personnel and others with a need for an understanding of Storage Networking will find this course extremely informative.

· Anyone involved in storage or data communications networking will understand the similarities and differences between these environments and will be in position to take on the challenges introduced by Storage Networking.

 

Prerequisites:

 

An understanding of current computer interfaces or networks is desirable, although not absolutely necessary.

 

Course Outline Overview:

 

Storage Network Overview

Storage Network Design

· Storage Networking Design Process Introduction

· Storage Networking Business &Technical Goals

· Storage Networking Information Flow Analysis

· Characterizing Storage Networking Infrastructures

· Topologies, Technologies and Products

· Logical and Physical Design

· Optimizing Storage Networks

· Document, Implement, Test

· Designing Fabrics

· High Availability Fabrics

· Design Summary

 

Performance Issues in Storage Networks:

· SAN Performance Prelude

· SAN Performance Flow Control

· Estimating Loop Performance

· Fabric Performance

Bus, Loop, Point-to-Point, Fabric and Hybrid considerations

Internal versus External SAN

High / Low Performance in SAN/NAS

 

Storage Networking Troubleshooting Approaches

State-of-the-art tools

Storage Network Futures

Course Summary

 

Course Outline Detail:

 

A.  SAN Design Process Introduction

 

· What is Design?

· SAN Design Process

· SAN Design Process: OSI

· SAN Design Process: Fibre Channel

· SAN Design Process: Elements

· SAN Design Steps: Top Down

 

B.  Storage Networking Overview

 

· Wires & Protocols

· Blocks Vs. Files

· What is a SAN

· SAN Benefits

· SAN Applications

· Storage Network Physical Transport Choices

· Network Components

 

C.  SAN Design Business & Technical Goals

 

· Analyze SAN Business Goals

· Typical Business Goals

· Customer's SAN Applications

· Constraints

· Misconceptions

· SAN Business Goals

· Business Goals Checklist

· Analyze SAN Technical Goals

· SAN RAS

· SAN Scalability

· SAN Performance

· SAN Performance/Efficiency

· SAN Security

· SAN Security Methods

· SAN Manageability FCAPS

· SAN Adaptability

· SAN Affordability

· SAN Technical Goals

· Technical Goals Checklist

· SAN Design Tradeoffs

 

D.  SAN Design Characterizing the Existing Infrastructure

 

· Examine Existing Design

· Check Health of the Existing Infrastructure

· Check Health: Tools

· Check Health of Existing Infrastructure

· Characterizing Traffic

· Characterizing Traffic: Models

· Characterizing Traffic: Flow Applications

· Characterizing Traffic: Flow Users

· Characterizing Traffic: Load

· Characterizing Traffic: QoS

· Characterizing Traffic

· Characterizing Checklist

 

E.  SAN Logical Design

 

· Logical Network Design

· Logical Design: Names & Addresses

· Logical Network Design

· SAN Topologies

· Point-to-Point

· Topologies: Arbitrated Loop

· Arbitrated Loop Performance Factors

· Arbitrated Loop Performance Factors

· Fabric Topology

· Fabric Performance Factors

· Fabric: Single Switch/Director

· Fabric: Multiple Switches

· Loop vs. Fabric

· Hybrid: Loop and Switch

· Loop and Switch

· Loop Switch: High Performance?

 

F.  Hierarchical Designs

 

· Hierarchical Design vs. "Connect-the Dots" Approach

· Hierarchical Design Mesh Topologies

· Hierarchical Design: Full Mesh Topologies

· Hierarchical Design: Full versus Partial Mesh

· Hierarchical Design: The Core Layer

· Hierarchical Design: The Distribution Layer

· Hierarchical Design: The Access Layer

· Hierarchical Design: Telephone Switching

· Hierarchical Design: Internet

· Hierarchical Design: SAN

· Hierarchical Design Guidelines

· Basic SAN

· Beyond Basic SANs Dept Class: Redundant Fabric

· Department Class SANs

· 20 Available Ports with 4 Switches

· 18 Available Ports with 6 Switches

· Department Class Backbone SAN

· Department Class SANs

· Enterprise Class Redundant Star SAN

· Enterprise Class SANs

· Enterprise Class Redundant Star-to-Star SAN

· Enterprise Class SANs

· Consolidated Core - Director

· Enterprise Server & Storage Connectivity

· SAN Topology Selection

· Topology Department Variations

 

G.  SAN Design Physical Design

 

· Create Physical Design

· Distance

· Flow Control

· Buffer-to-buffer credit

· End-to-end credit

· Distance: Footprint

· Fibre Channel Clustering

· Storage Clustering

· Compute server clustering considerations

· Clustering factors

· Cabling

 

H.  Protocol Optimization

 

· Protocol Optimization: Data Rate Bus Factors

· Protocol Optimization: Parallel SCSI Overhead

· Protocol Optimization: Serial Communication

· Equipment Selection

· Equipment Selection: HBA

· Equipment Selection: System Bus Architecture

· Equipment Selection:  Media Interface and Protocol Engines

· Equipment Selection: Switches

· Automatic Fail-over?

· Persistent Binding – Path Consistency

· Equipment Selection: Bridges

· Equipment Selection: Servers & Storage

· SAN Design Optimization

· Optimize SAN Design

· Optimize SAN Performance

· Optimize SAN Performance

· Amdahl's Law

· Adjust Timer Values

· Optimize Block Size

· Optimize Cache

· Maximize Bandwidth: Aggregation

· Maximize Bandwidth: Utilization

· Optimize Port Consolidation

· Capacity Planning

· Capacity Planning: Distribute File System Activity

· Capacity Planning: Increase SCSI Command Queuing

· Capacity Planning: Analyze Flows "Forgetting about" SANs

· Capacity Planning: Scaling & Disk Growth

 

I.  SAN Document, Implement & Test

 

· Document SAN Design

· SAN Design Document

· SAN Design Document

· SAN Implementation

· SAN Implementation Examples

· SAN Test

 

J.  SAN Design Summary

 

K.  Designing a Fabric

 

· Designing the Fabric

· Single-Switch Fabrics

· Multi-Switch Fabrics

· Multi-Switch Fabric

· Multi-Switch Configurations

· String Configuration

· Linear String Configuration

· Ring Configuration

· Tree Configuration

· Redundant Tree Configuration

· Fully Connected Mesh

· 4 Switch Fully Meshed Fabric

· 6 Switch Fully Meshed Fabric

· Partially Connected Mesh

· Partially Connected Meshed Fabric

· Designing a Fabric End

· High-Availability Fabrics

· RAS

· Reliability

· Failure Rates vs. Time

· Using Redundancy

· Redundant Array of Disks (RAID)

· Redundant Controllers

· High-Availability Fabrics

· Redundant Links Between Switches

· Redundant Fabrics

· Redundant Switches?

· Redundant Fabric Rings

· Concurrent Maintenance

· Disaster Recovery

 

L.  SAN Performance Prelude

 

· Industry Quotes

· Moore’s Law

· Growth in Disk Capacity

· Rethinking the I/O Interface

· Performance vs. Distance

· System Bus

· I/O Interfaces

· Input/Output (I/O) Interfaces

· Local Area Networks (LAN)

· LAN Interface

· Wide Area Networks (WAN)

· The Picture is Changing

· Boundaries Are Blurring

· Getting to the Data

· The Road to the Data

· Computing Paradigms

· Centralized Computing

· Centralized Computing

· Centralized User Access

· Single-User System Dataflow

· Multi-User System Dataflow

· Direct Attached Storage

· Channel Attached Storage

· Channel Design Objectives

· Channel Attached Storage

· Sharing Storage Via a Channel

· External Storage Controller

· External RAID Controller

· Channel Limitations

· The Local Area Network

· The LAN Environment

· LAN Data Access

· LAN Client-Server Dataflow

· The LAN Client

· The LAN Server

· Network File Systems

· LAN Server with RAID

· Network File Sharing

· Network Attached Storage

· Network Attached Device

· Network Attached Device

· The Network Computer (NC)

· Network Computing

· Network Computing Access

· Network Computer Dataflow

· LAN Backup Issues

· LAN Limitations

· The New Paradigm

· Storage Area Network (SAN)

· Storage Area Network

· Speed of Signals

 

M.  Performance Flow Control

 

· FC-2 Data Rate Calculation – Class 2

· FC-2 Data Rate Calculation – Class 3

· Flow Control

· Transmit and Receive Buffers

· Credit

· Obtaining Credit

· Credit and Frame Streaming

· Credit and Long Links

· Frame Transmission and Credit

· Timing Example for Credit=1

· Credit Computation

· Estimate Credit Procedure

· Estimate Credit Procedure

· Buffer-to-Buffer Flow Control

· Topology Dependencies

· Alternate BB_Credit

· BB_Credit Recovery

· Virtual Circuit Flow Control

· End-to-End Credit

· Flow Control

· Network Bisection Bandwidth

 

N.  Estimating Loop Performance

 

· Arbitrated Loop Performance

· Loop Bisection Bandwidth

· FCP Information Units

· Loop Tenancies

· Control Command

· Control Command

· Write Command Tenancies

· 4-Tenancy Write, BB=0

· Read Command Tenancies

· Key Performance Factors

· Loop Round-Trip Time

· Loop Tenancies

· Half-Duplex Ports

· Login BB_Credit

· Opportunistic Transfers

· Performance Example

· Loop Round Trip Time

· Command Frame, BB=0

· Xfer_Rdy Frame, BB=0

· 8k Data Transfer, BB=0

· Status Frame, BB=0

· Effective Data Rate

· Input/Outputs Per Second

· Service Rate

· Loop Utilization

· Arbitration Wait Time

· Arbitration Queue Depth

· Effect of Utilization

· Arbitrated Loop Performance Estimate

 

O.  Fabric Switch Performance

 

· FC Switch Bisection Bandwidth

· FC Switch Bisection Bandwidth

· Inter-Switch Link Contention

· Multi-Pathing

· Path Analysis

· Interrupts

· Multi-Pathing

· Load Sharing vs. Load Balancing

· Load Trunking

· Fabric Inter-Switch Link

· Fabric Inter-Switch Links

· Round Robin

· Load Sharing – Round Robin

· Fabric Inter-Switch Links

· Trunking

· Trunking: Stripe the FC Transmission Word

 

P.  Storage Networking Troubleshooting

 

· A Structured Approach to Finding and Fixing Problems

· SAN Tips, Tricks and Traps

· Bad Connections

· Microbends

· Macro Bends Dispersion

· Too Many Patch Panel Connections

· Fibre Optic Handling

· Fibre Optic Installation

· Electrical Links

· Interoperability

· Procure Appropriately

· Arbitrated Loop vs. Fabric

· Data Sharing

· Things to Watch

· Latency

· Over-Subscription

· Lack of bandwidth

· Too much bandwidth

· Interoperability