Brocade: Storage Networking For the Virtual Enterprise

1 © Copyright 2012 EMC Corporation. All rights reserved.

Storage Networking For the Virtual Enterprise AJ Casamento


Legal Disclaimer All or some of the products detailed in this presentation may still be under development and certain specifications, including but not limited to, release dates, prices, and product features, may change. The products may not function as intended and a production version of the products may never be released. Even if a production version is released, it may be materially different from the pre-release version discussed in this presentation.

Nothing in this presentation shall be deemed to create a warranty of any kind, either express or implied, statutory or otherwise, including but not limited to, any implied warranties of merchantability, fitness for a particular purpose, or non-infringement of third-party rights with respect to any products and services referenced herein.

Brocade, Brocade Assurance, the B-wing symbol, DCX, Fabric OS, MLX, SAN Health, VCS, and VDX are registered trademarks, and AnyIO, Brocade One, CloudPlex, Effortless Networking, ICX, NET Health, OpenScript, and The Effortless Network are trademarks of Brocade Communications Systems, Inc., in the United States and/or in other countries. Other brands, products, or service names mentioned may be trademarks of their respective owners.


Abstract Storage Networking For the Virtual Enterprise

This Session will review the latest in storage networking protocols for EMC storage platforms. Topics include new architecture for scaling SAN fabrics, new features for enhancing SAN metro connectivity, and powerful new capabilities for enhancing SAN Management. We will also review Ethernet Fabric architectures for FCOE, iSCSI and NAS.

At the end of this session you will be able to: – Understand how to deploy large SAN Fabrics using multiple

directors without sacrificing server and storage ports. – Learn about the new SAN capabilities that will significantly

improve Metro connectivity over DWDM or dark fiber. – Understand how Ethernet Fabrics provide performance, scale and

ease of manageability compared to traditional Ethernet hierarchical topologies.


From the Data Center to Private Cloud Storage


Business Objectives

Non-stop access to information – Fast, reliable

Move at speed of business – Scale with growth and application

needs – Deliver new services on demand – Rapidly adjust to changing conditions

Reduce costs – CapEx and OpEx – Maximize investments – Gain efficiencies


Drivers in Data Center Change

Demand for greater business agility--to deliver information to any device on any network at any time

Need for on-demand availability of resources to support "pay as you grow" model

Increased server virtualization

Escalation of bandwidth from 1 Gigabit Ethernet (GbE) to 10 GbE and higher due to unpredictable workload requirements and location

2008 2009 2010

Virtualization Growth

Newly deployed Virtual Machines (VMs)

Newly deployed physical hosts

The Tipping Point

Source: IDC

By 2018 86% of server workloads will be on VMs

22 BILLION INTERNET-CONNECTED

DEVICES BY 2020 Source: IMS Research, 2010


Evolution of the Data Center

2002 2008 2011 2014

Consolidation

Virtualization

Private Cloud

Hybrid Cloud

Private Clouds

Public Clouds

Hybrid Cloud


Transition to Private Cloud Infrastructure

Traditional Data Centers Cloud-Enabled Data Centers

Remote Data Centers

Private cloud attributes • Highly virtualized pools of compute, storage, and network resources • On-demand, fast provisioning of application resources • Lower capital and operational costs, higher asset utilization • Automated management and orchestration

Remote Data Centers


Private Cloud Storage Requirements

Maximize infrastructure consolidation and virtualization Simplify distance connectivity

between data centers Adapt to rapid growth and change

in storage infrastructure Automate management and

integrate with orchestration frameworks Increase storage performance for

emerging workloads Central and remote data centers connected over a private cloud


DATA CENTER FABRICS

Ethernet Fabrics

Fibre Channel Fabrics


15 years of industry leadership

in performance, reliability, scalability, and energy efficiency 90% of Global 1000 data centers

324 patents issued worldwide

and 328 patents pending Consistently first to market

BROCADE FIBRE

CHANNEL FABRICS


Maximize infrastructure consolidation and virtualization


Networks that adapts to your business Reduce network complexity, management, and costs

Simplifies scale-out network design to reduce network complexity, management, and costs Maximizes overall port density and

space utilization for cost savings through massive consolidation of legacy SANs Features industry-leading

performance to support traffic growth and increased requirements for virtualized environments


Brocade DCX 8510 Backbone Family Proven scalability, performance, and reliability

Brocade DCX 8510-8 • Up to 384 16 Gbps Fibre Channel ports

or up to 512 8 Gbps Fibre Channel ports • Large enterprise fabric core

Brocade DCX 8510-4 • Up to 192 16 Gbps Fibre Channel ports or

up to 256 8 Gbps Fibre Channel ports • Midsize enterprise fabric core • Large enterprise edge or application

engine Two backbone models


Scale-Out Optical ICLs Backbone scalability, performance, and reliability Maximizes performance and scalability

– Connect up to ten chassis (core-edge) – 32 × ICL ports per Brocade DCX 8510-8 chassis – Each ICL port delivers 64 Gbps bandwidth – Up to 2 Tbps ICL bandwidth (four times existing) – Up to 50 meter cables for flexible configuration

Minimizes latency between chassis – Does not count as a hop for FICON environments

Maximizes load balancing and availability – Frame-based trunking is automatically enabled

between up to four ICLs Links – DPS distributes exchanges across all frame

trunks

New core routing blades with optical ICL ports

New ICL optics QSFP (4×16 Gbps)

8510-8 Core 8510-4 Core


Scale-Out Fabrics: 9 Chassis Full-Mesh with ICLs Simple, high performance mesh topology

Simpler, flatter, low-latency chassis topologies

– Nine chassis active-active mesh

Higher port density and bandwidth

– Supports up to 3400+ 16 Gbps ports or 4600+ 8 Gbps ports with no Inter-Switch Links

– Delivers nearly 74 Tbps of aggregate bandwidth in 9 domains

Fewer chassis, fewer cables – Mesh with ICLs: 9 chassis, 144

cables – Core/Edge with ISLs: 12 chassis

1458 cables


Scale-Out Fabrics: 10 Chassis Core-Edge with ICLs Scalable, cost-effective core-edge topology

ICLs enable simple, scalable core-edge topologies

– Ten chassis core-edge – DCX 8510-8 core, DCX 8510-4 edge

Higher port density and bandwidth

– Supports up to 2300+ 16 Gbps ports or 3000+ 8 Gbps ports with no Inter-Switch Links (ISLs)

– Delivers 49 Tbps of aggregate bandwidth in 10 domains

Fewer chassis, fewer cables – Core-edge with ICLs: 10 chassis, 64

cables – Core/Edge with ISLs: 12 chassis

1458 cable


Simplify distance connectivity between data centers


Networks that keeps running Integrated metro and global SAN extension

Leverages Fibre Channel and FICON technologies that are proven in 90 percent of Fortune 1000 data centers Enables flexible, high-speed

replication solutions over metro or WAN links with native 10 Gbps Fibre Channel or 10 GbE FCIP extension support Features enhanced diagnostics and

RAS functionality to help minimize disruptions and downtime


ISL: In-flight Compression & Encryption Move More Data and Encrypt on the fly

Storage traffic gets compressed on ISL and gets uncompressed at the receiving switch

Provides up to 2:1 compression and uses Brocade LZO algorithm

Up to 4 ports per DCX 8510 blade, 2 ports per 6510 switch

Encryption works the same way

Uncompressed

Compressed

Critical Data

Critical Data Critical Data

Critical Data


ISL: End-to-End Optics & Link Validation Ensure fabric link level integrity

Identify and isolate optics and cable problems faster

Reduce fabric deployment and diagnostic times

Non-intrusively verify transceiver and cable health

Provide granular latency and distance measurement for buffer credit assignment

Test electrical and optical transceiver components

– Monitor transceiver health – Conduct cable health check – Set alerts for digital diagnostics

16 Gbps SAN Core

16 Gbps SAN Edge


FCIP: Fast Write to Mitigate Latency Go Farther and Faster

0

5000

10000

15000

20000

25000

0 5 10

DVD Store OPM Fast Write Off

DVD Store OPM Fast Write On

RTT (ms) DVD Store OPM Fast Write Off

DVD Store OPM Fast Write On

5 10561 17530 10 6747 13496

10ms, Fast Write =on, OPM =13,496

5ms, Fast Write = off, OPM =10,561

Enabling FCIP Fast Write, yields 30% more I/O and 2X the Distance (at 10ms RTT)


FCIP: Non-Disruptive Link Recovery L3 Lossless Link Loss (LLL) using FCIP Trunking

1 3

4

X X X Circuit 1 with TCP session

Circuit 2 with TCP session

FCIP Batches

FCIP Batches

3?

Supervisor Session Data Received

3 2

Supervisor Session • FCIP Trunk appears as a single logical ISL • Encapsulates the circuits’ TCP sessions within a FCIP Trunk • Provides granular load-balancing • Retransmits lost frames due to circuit loss • Ensures FC frames are delivered in order (IOD) • Supports FastWrite, FICON and Tape Pipelining


Automate management and integrate with orchestration frameworks


Networks that are simple and easy to manage Enhanced diagnostics, monitoring, and management

Simplifies and centralizes management through Brocade Network Advisor, reducing operational costs and complexity Delivers critical monitoring and

integrated diagnostic capabilities for optimal SAN performance and reliability Supports partitioning for Multi-

tenancy deployments


RAS: Bottleneck Detection

Identifies and alerts administrators to bottlenecks that can degrade application performance

– Detects bottlenecks caused by slow drain devices

– Bottleneck detection ▪ E_, EX_, and F_Ports

Accelerates problem detection and diagnosis to minimize performance degradation

Latency Bottleneck Monitor, F_Port

Slow-drain Device

Congestion Bottleneck Monitor, E_Port

Congestion Bottleneck Monitor , E_Port

ISL Congestion

Normal Traffic

Congested Traffic


Multi-Tenancy Support: Virtual Fabrics Logically partition a physical SAN

Partition Brocade DCX 8510 into logical switches

Connect logical and “unaware” switches to form logical fabrics

Isolate and manage by application, business group, customer, or traffic type (Fibre Channel, FICON)

Partitioned Brocade DCX 8510-8 using Virtual Fabrics

Logical Switch 1

Logical Switch 2

1

Logical Switch 3

2

Logical Fabric 1

Logical Fabric 2

Logical Fabric 3

“Unaware” Brocade DCX 8 Gbps

LS 2

1

LS 1

2


Increase storage performance for emerging workloads


Emerging and Evolving Workloads I/O intensive applications Unmatched I/O performance

– 500,000+ IOPS per HBA – 700 ns latency, 420 million fps per

switching ASIC – Eliminates network bottleneck with SSD

storage

Supports (VDI) deployments for more than 2000 clients

– Scalable, high-performance storage for boot/login storm and peak loads

– Maximum virtual desktop density per server

Thin Clients

~1000 VDIs POD-1

~1000 VDIs POD-X

Brocade SAN

Solid State Drive (SSD) Array

Supports evolving tier one workloads

Lowest latency for high-transaction storage

Maximum I/O and bandwidth for application performance


Server Virtualization Optimization 16 Gbps fabric adapts to increased workloads

Optimized performance for high-density VM deployments

– 2X the bandwidth of 8 Gbps – Faster I/O performance

Integrated VM management – Plug-in for VMware vCenter

/Microsoft System Center – VM-to-storage port monitoring

Improved flexibility for virtualization

– End-to-end Quality of Service (QoS) for VM-level traffic prioritization

Hypervisor Management plug-in


Connectrix Manager Converged Network Edition Ease of use

– Group management of directors and switches

– Automation of manual and repetitive Tasks

Configuration Management – Wizards-based zoning

configuration, including LSAN zoning and zoning reports

– Simple wizards to configure FC, FICON, FCIP, encryption, and HBAs/CNAs

– SAN diagnostics, monitoring, and bottleneck detection

– Simplified management of virtual fabrics

Image and change management Monitor configuration changes Image management, including

snapshot Back up switch configuration


Cloud-Optimized SAN Platforms

Rely on a network that keeps running, no matter what

Build a network that adapts to your business

Manage your network simply and efficiently

8-slot and 4-slot Brocade DCX 8510

48-port 1U Brocade 6510

Single and dual-port Brocade 1860 Fabric

Adapters

24-port 1U Brocade 6505

Fibre Channel is the foundation for private and public cloud storage


DATA CENTER FABRICS

Ethernet Fabrics

Fibre Channel Fabrics


Resilient architecture for today’s consolidated environments

Flexible topology to meet application and storage requirements

Massive scalability while remaining elastic

Inherently flat architecture for virtualization and clouds

WHAT IS AN ETHERNET FABRIC?


–Albert Einstein

“Everything should be as simple as possible, but no simpler.”


Brocade VCS Ethernet Fabric Scalable Flat layer 2

Core

Ag

greg

atio

n

Servers with 1 or 10 Gbps Connections

Edge

Scalability

Acce

ss

Ethernet Fabric Architecture

Classic Hierarchical Ethernet Architecture


Brocade VCS Ethernet Fabric Auto-Healing and Improved Utilization

NAS iSCSI FCoE

Automatically reroutes traffic during failure

Active—active resiliency and load balancing

Arbitrary topologies

Multi-path, deterministic

Low latency

Lossless, convergence capable

VCS Fabric


Brocade VCS Ethernet Fabric Operational Automation and Simplification

• Self forming fabric

• Managed as a single switch

• Logically collapses network layers

• Auto-discovery and propagation of device information

• Automatic migration of port profiles


Services in Classic Architecture L

3 Co

re

L3 A

ggre

gatio

n La

yer 2

La

yer 2

Se

rver

Rac

ks

CHALLENGES

Utilizes expensive L3 aggregation ports

Bandwidth constraints at aggregation layer

Manual management of service policies in VM move scenarios

Rigid network design forces services to be deployed at specific points in the network, limiting elasticity and resource pooling


Brocade VCS Ethernet Fabric Dynamic Services Insertion

L3

Core

La

yer 2

/ La

yer 3

La

yer 2

Se

rver

Rac

ks

BENEFITS Fewer service appliance

are needed

Enables virtual appliances

Eliminates bandwidth constraints

Applications placed and moved anywhere in the fabric with instant access to necessary services

No need for utilizing more expensive aggregation/core ports for services


Manual VM Migration

Administration of today’s environments is time consuming; dampens adoption of cloud architectures

Elasticity is reduced; compute and storage assets restricted from forming pools of resources


Brocade VCS Ethernet Fabric VM Aware Network Automation

No need for manual configuration of MAC addresses and port profiles; Less error-prone

Minimizes procedural delays between server and network IT teams

Eases configuration of multiple VCS fabrics

Protection against VM/MAC Spoofing via secure vCenter Communication

Brocade Network Advisor

NAS iSCSI FCoE FC iSCSI NAS

vCenter


Resilient architecture for today’s consolidated environments

Flexible topology to meet application and storage requirements

Massive scalability while remaining elastic

Inherently flat architecture for virtualization and clouds

BROCADE VCS

FABRICS


Transition to Private Cloud Architectures

Private cloud architectures – Simplify infrastructure – Become more agile – Realize CapEx and OpEx savings

Brocade fabric-based solutions lead the industry

– Essential foundation for cloud-optimized networks

– Simple, flat, and high-performance – Highly resilient and scalable networks

Deliver the full benefits of virtualization in private clouds


Thank You

Technology

Brocade: Storage Networking For the Virtual Enterprise