© 2013 ibm corporation networking fundamentals andy wright global field marketing, ibm system...

© 2013 IBM Corporation

Networking FundamentalsAndy WrightGlobal Field Marketing, IBM System Networking

[email protected]

IBM TGVL: System Networking FoundationsJune 2013


System Networking Top Gun

Instructor Bio: Andy Wright

Experience:

Over thirty years of experience in Information Technology Services Sales and Marketing.

•2013-Present Channel and Sales Enablement – IBM System Networking

•2004-2013 Product Marketing Director – Global Knowledge

•2002-2004 Sales Manager – Channel Advisor

•2000-2002 Marketing Director – Static Control Components

•1981-2000 IBM (Netfinity/System x Marketing, ThinkPad Product Marketing,

and Sales positions in CA and MN, IT Programming)

Education:

•Bachelor of Science, Business Admin Management & Computer Science, University of

Wisconsin - Platteville

Hobbies:

• Beach, Civic Organizations, Travel



Agenda

Basics – what is a network, components, terminology

OSI model & layers

Virtualization & networking

Storage & network convergence



What Is A Network?

Physical and logical interconnection of “internet aware” devices

Networks provide an infrastructure to share resources and information

A network is a resource, or group of resources, that share common communication infrastructure to distribute data

Networks can scale greatly in size, ranging from two to three machines, to millions of interconnected networks

Clients, servers, network interface



Network Interface and Speeds

Devices have network interfaces – port or nic (network interface card) Ethernet is the standard Speeds at which computers and servers can access data over the network is

determined by the speed of the ‘wire’ they connect to These ‘wire’ or ‘media’ speeds are rapidly increasing to mind bogglingly fast data

rates

– 1 Gb speeds over Ethernet has been the standard

– 10 Gb becoming more persuasive

– 40 Gb products exist

– 100 Gb is coming Ethernet has become more reliable with advent of Lossless

– Lossless Ethernet: enhancement to Ethernet for data center environments• Data Center Bridging (DCB) indicates lossless Ethernet – used by IEEE standards

group. • Convergence Enhanced Ethernet (CEE) – IBM’s trade marked term• Data Center Ethernet (DCE) – Cisco’s trade marked term



Parts of a Network

Node - a device (servers, storage, computers, etc) connected to the network Switch – connects devices to allow devices to communicate to each other Router – helps pass data to the correct device

Switch RouterNode



Access Devices (or Endpoint Devices)

Devices that actually utilize network resources In the data center, these are typically servers, or storage devices In the office, this includes anything with an Ethernet port, including PCs,

IP phones, POS systems, and virtually anything else that might need to talk to a network

Access devices need to connect, or “uplink” with the rest of the network. This is done by connecting to a switch



Routers

Act like traffic GPS units Responsible for understanding where traffic needs

to be moved across the network Also responsible for knowing whether it is

appropriate to be sending certain devices’ traffic to other devices across the network

Utilize technologies to allow multiple users to share a single internet connection

Analogy: Real-time translator. It will know that if the finance department speaks English, in order to talk to the human resources department, it will need to translate to French in real time. The router will also know that the employees in marketing probably shouldn’t need to talk to the payroll department



Ethernet Switches

Act as the “skeleton” of the network – provide the framework both physically and logically to connect endpoint devices to the network. No switches, no network!

Unlike other network devices, switches typically have a high port density, meaning, a larger port count (often 24 to 48+ ports!)

Have transitioned from “connectivity” boxes to complex devices capable of performing all of the aforementioned roles, and sometimes, more

A lot of players in this market, including IBM’s System Networking portfolio, in addition to products from Cisco, HP, Juniper and many, many others



Local and Wide Area Networks

Computer networks allow applications operating on various computers to communicate with one another.

New YorkLAN

LondonLAN

WAN

Networking Standards

Ex IEEE

Server 1

Server 2

Vendor Interoperability• Competition - connect diff vendors

• Ease of Networking



Networking OSI Reference Model

11

Layer 7 Application

Layer 6 Presentation

Layer 5 SessionLayer 4 TransportLayer 3 NetworkLayer 2 Data LinkLayer 1 Physical • Transmission method of bits

•Transmission across physical link (LAN/WAN)

•Addressing and packet transmission (best path)

•Data tracking (End-to-end connection)

•Job Management tracking Inter-host communication

•Data representation (Encoding Language)

•Network processes to applications

Destination Source VLAN Type / Length Data CRC

Layers 2 – 3 (Know from & where to) Layers 4+ (what’s inside)



Network Layers Continued

The key differentiation between most switches is their degree of intelligence

Switches typically fall under the category of: Layer 2, Layer 2/3, and Layer 4+ So, what does this mean? Let’s quickly discuss the concept of a frame

Destination Source VLAN Type / Length Data CRC

Layers 2 - 3 Layers 4+



Network Layers – The Analogy

Frames are little packages wrapped up inside each other When network devices can understand what’s wrapped up deep inside a

frame, they’re able to make intelligent decisions about how to interact with frames

Layer 2/3 switches understand where the frames come from, and where they have to go

Layer 4+ switches understand what’s inside a frame, and can actually interact with frames to perform intelligent actions like reply to a system’s request for information before it is actually processed by the destination system

Allows for intelligent load balancing and other complex operations based on content rather than source or destination



Layer 1: Physical Layer

14

Defines the physical characteristics of the network such as connections Physical media type does not dictate the type of internet traffic that can flow across it.

Copper RJ45, DAC SFP+ or Fibre, all can carry “Ethernet” traffic Whether travelling by train, bus, or car, as long as the objective is to get from point A to

point B, the method travelled is irrelevant

Ethernet standards - IEEE 802.3

Support different distances – Short Range 100’s meters, Long Range 10 KM

Pass-through would be an example (EN4091)

Copper RJ45(CAT5, CAT6, CAT6e)

SFP+ DAC(Passive / Active)

Fibre Optics(Ex. LC/LC)



Layer 2: Data Link Layer

15

Provides reliable transit of data across a physical link Defines physical addressing, network topology, and is also concerned

with error notification, sequencing of frames and flow control. Consists of:

– MAC layer - Provides unique address for an NIC– LCC (Logical Link Control) layer

Examples of Layer 2 switches– All the Ethernet devices in IBM portfolio support Layer 2 (except

Pass-thru)– Layer 2 offerings:

• All RackSwitch offerings• Embedded - EN4093/R, CN4093

• Basic L2 only: Easy Connect Mode & SI4093



Layer 3: Network Layer Provides connectivity and network path selection between two networks. Required when going across domains or VLAN’s "What's the difference between a Layer 3 switch and a router?"

– Layer 3 switches have optimized hardware passing data traffic as fast as Layer 2 switches our usually faster because it is built on switching HW.

– Layer 3 switches make decisions regarding how to transmit traffic like a router does.

Examples:

– All IBM RackSwitch and embedded switches

– Reduce need for external switches & ports

– Can help reduce costs• Acquisition costs

• Operating costs – power, cooling & IT resources

– Can improve availability & security



Campus

Access Layer

Aggregation Layer

Core Layer

Early Ethernet Campus Evolution

In the beginning, Ethernet was used to interconnect stations (e.g. dumb terminals), initially through repeater & hub topologies,

Then eventually through switched topologies. Ethernet campuses evolved into a structured

network typically divided into a Core, Service (e.g. firewall), Aggregation and Access Layer.

– Typically: core, services, aggregation & access planes.

– Traffic is mostly North-South(directed outside campus).

– To avoid spanning tree problems, campus networks typically are divided at access.

The industry liked the tree structure & applied to DC

WANWAN

5%5%

95%

95%

Lay

er3



Laye

r2SAN

Ethernet in the Data Center

But the Data Center is differentTraffic patterns are East-West* Large layer-2 domains needed for clustering and Virtual Machine mobilityHas different fabric performance needs

– Lossless traffic for storage

– Low latency & high bandwidth for clustersEvolved into a virtual compute model, with different demands:

– From static workloads to dynamic workloads

to multi-tenant & dynamic

Data Center

Access Layer

Core Layer

Aggregation Layer

> 75%> 75%

< 2

5%<

25%

WANWAN



Virtualization and Networking Gartner states that virtualization and consolidation are the two largest movements

occurring in the data centre While the greatest degree of focus has been on the virtualization of “traditional”

computing systems like servers, virtual network technology is becoming a huge market

Virtualization vendors like VMware and Xen have products to meet this need, however 3rd party networking vendors like IBM have also developed incredibly elegant and sophisticated technology to enable administrators to virtualize network equipment

Virtualization technologies like IBM’s Virtual Fabric, VMready and SDN VE are changing the way IT personnel are thinking about network technologies

The technologies currently available to facilitate network virtualization are discussed in greater detail in the Networking Trends education module



Virtualization: Virtualizing Network Components

While a more technically involved concept, it is important to understand that while virtual hardware is logical, it is ultimately dependent on physical hardware in order to actually work in a way that is useful across a network

With industry buzzwords like “virtual switches, virtual NICs and virtual fabrics” being discussed, understanding that networking vendors like IBM have a wealth of technologies available to meet each of these unique requirements that work well with other vendors – this is a huge advantage.



Standards will impact future Data Center

Virtualization IEEE 802.1Qbg

– Server-network edge virtualization – Uniform view of VMs in the

hypervisors and the network– Visibility of VM traffic in the network– Automatic migration of port profiles

IEEE 802.1Qbh/r– TAG added in Hardware to the

Ethernet frame.

OpenFlowSilos – app specific

WEBAP

DB

WEBAP

DB

WEBAP

DB

WEBAP

DB

IP NWIP NW▌ Inefficient

▌ Complex

▌ Slow response

▌ Static

▌ Vendor lock

Optimization software control

WEBAP

DBWEBAPDB

WEBAPDB

WEBAPDB

IP NWIP NW▌ Optimized

▌ Simple

▌ Agile

▌ Dynamic

▌ Low cost



SAN – The Storage Area Network

Servers need storage! In larger scale networks, this typically means servers are attached to storage systems over the SAN on separate physical infrastructure

This technology is typically driven over optical FibreChannel connections, however FibreChannel does run on copper cable as well

The storage networks have traditionally been kept separate from data networks because storage technology is delivered to servers at a “block by block” level. Basically, the electronic representation of physical sectors on a hard drive is transmitted back and forth between servers and storage. This means that storage networks must be very high speed, and incredibly reliable. Data networks are “lossy” networks, which is unacceptable for storage

FC Switch

WWN

FCoE to FC Gateway

10G Loss Less Ethernet2/4/8G Fiber Channel

F-port N-portWWN WWN MAC

MAC

NPIV F-port

CNA



SAN – How We Talk to Storage

Similarly to data networks, storage networks still need devices like switches to

extend the availability of their resources. IBM System Networking has a large

portfolio of SAN systems from both Brocade and Cisco

As will be discussed in greater detail shortly, storage networks cannot directly talk

to data networks in most cases. In order to address this challenge, there are

devices that enable “data centre bridging”, which is essentially using a gateway

device to literally bridge the connection between the data and storage network

IBM System Networking offers a portfolio of 16 SAN devices



Network Convergence: Overview

Historically, different network services required different communications infrastructure

Telephones ran on telephone networks, storage was accessed over storage networks and data was contained within data networks

This is no longer the case! As the performance of networks increases, and vendors are able to produce

technology capable of meeting the needs of multiple technologies sharing a common ‘wire’, there is, and will continue to be a major shift away from physically decoupled networks to physically shared but logically segregated networks

Network convergence is rapidly becoming the biggest trend in the networking world



Gateway

iSCSI Storage

Array

FCoE Storage Array

NFS/CIFS NAS FC Arrays

Fibre Channel

Lossless

Ethernet

Gateway

This educational material is intended for your use in selling. It is NOT a deliverable for your clients.25

Network Convergence: What Is It?



Network Convergence: FibreChannel & Storage Area Networks Typically in enterprise environments, storage has been attached to the network via an

independent, physically decoupled infrastructure known as the SAN, or storage area network The de-facto method of attaching servers to storage has been to leverage the FibreChannel

protocol, which uses optical connections and a specific ‘protocol’ or ‘language’ to facilitate communication between servers and storage systems

IBM markets a number of SAN-specific products including FibreChannel switches, as FC is still the dominant product in this market

FC systems have also traditionally operated at faster speeds than Ethernet technology

FibreChannel does have several drawbacks however, mainly the cost associated with maintaining and procuring independent infrastructure just for storage

As will be discussed in the following slides, convergence is rapidly negating the need for physically independent networks



Network Convergence: The Challenges of Current Technologies

There are two major roadblocks impeding the rate at which convergence is being adopted

– Many networks still operate with gigabit Ethernet technology. Unfortunately this speed is insufficient to properly trunk high bandwidth storage applications

– The second major challenge is that Ethernet technology was not lossless meaning that the technology is designed with the assumption that frames will be dropped by networking devices if certain conditions like excess capacity or network problems occur

Storage data is incredibly sensitive to data loss and latency and thus Ethernet as currently implemented is not best suited for many mission-critical storage applications

There is technology currently being standardized to address these challenges



Network Convergence: Leveraging New Technologies

10 gigabit per second infrastructure is now becoming the standard in data centres which greatly increases the suitability for Ethernet in storage applications

40 gigabit per second technology is already available (with limited implementation, however) and standards bodies are working 100+ gigabit per second technology

Lossless communications technologies such as CEE, or Converged Enhanced Ethernet eliminate the issues associated with network packet loss by implementing QoS

FCoE, or FibreChannel over Ethernet allows FibreChannel “frames” to be encapsulated, or “packaged” inside of Ethernet frames permitting traditional network equipment to understand how to talk to storage equipment

iSCSI and NAS will allow administrators to utilize the same technological concepts to increase availability and access to their network attached storage

Other technologies that will be able to leverage lossless communications including high performance computing clusters and computational systems



IBM System Networking - Overview

Mission: Help IBM deliver on the promise of the best “System’s Company” by providing the best server-to-server, server-to-storage, storage-to-storage, and system-to-system interconnect

Integral part of many IBM System and Solution offerings – Optimization, automation and integration

Deliver lower latency, loss less, low cost and low power technology

Built on industry standards & designed for future

Seamless integration with Cisco, Juniper and others

Critical in fighting off the competition, increase revenue & profit

Storage Software

Servers

Networking



ありがとうございました

MerciGrazie

Gracias

Obrigado

Danke

Japanese

Hebrew

Thank YouEnglish

French

Russian

German

Italian

Spanish

Brazilian Portuguese

Hindi

Tamil

Korean

Thai

Simplified Chinese

ArabicJapaneseJapanese

Dziękuję

Korean

Polish

Terima Kasih Malay


Learning Points

This educational material is intended for your use in selling. It is NOT a deliverable for your clients.




Networking is critical to clients business A network is the physical and logical connect of devices A network is a resource that shares a common communication infrastructure

to distribute data Ethernet networks speed and reliability has increased tremendously over

the years Servers and storage are nodes also know as access or endpoint devices

that utilize network resources Ethernet switches provide the framework both physically and logically to

connect endpoint devices to the network. Switches typically have high port density.

32

Learning Points – Part 1




OSI Reference Model is made up of seven layers: – Layer 7 Application: Network processes to applications– Layer 6 Presentation: Data representation (Encoding)– Layer 5 Session: Job Management track– Layer 4 Transport: Data Tracking– Layer 3 Network: Addressing and packet transmission (best path)– Layer 2 Data Link: Transmission across physical link– Layer 1 Physical: Transmission method

Switches typically fall under the category of Layer 2, Layer 2/3 and Layer 4+ Frames are packages containing information about the destination, the source,

VLAN, Type/Length, Data and CRC (for validation and error checking) Layer 2/3 switches understand where the frames come from and where they are

supposed to go to Layer 4+ switches understand what’s inside a frame and can interact with the

frame to perform actions like reply to a system’s request for information before it’s actually processed by the destination system

33





Layer 1 – the physical layer–Defines the physical characteristics like the connection – Can be copper, fibre optical, or SFP+ DAC. All can carry Ethernet traffic–The physical cable varies in distances it can carry signals

Layer 2 – the data link layer– provides for the reliable transit of data across a physical link–Defines the physical addressing, network topology and error notification–Consists of MAC layer – provides unique address for a network interface card (nic) and Logical Link Control layer

Layer 3 - Network Layer–Provides the connectivity and network path selection between two networks–Required when going across domains or VLANs.

34





Early Ethernet Campus – used to connect dumb terminals through hubs and repeaters. Eventually switches were used and almost all traffic flowed North – South through a core switch. Ethernet in Data Center – traffic patterns are East – West server to server, virtual machine to virtual machine. Virtualization has made significant changes to servers and storage. Now network virtualization is gaining popularity. Storage has typically used FibreChannel connections With the advent of more reliable (lossless) and higher performance, Ethernet is gaining acceptance as a viable alternative to FC. Network convergence is becoming one of the biggest trends in the networking world.

35





Disclaimers and Trademarks




8 IBM Corporation 1994-2013. All rights reserved.References in this document to IBM products or services do not imply that IBM intends to make them available in every country.

Trademarks of International Business Machines Corporation in the United States, other countries, or both can be found on the World Wide Web at http://www.ibm.com/legal/copytrade.shtml.

Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries.Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both.Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both.UNIX is a registered trademark of The Open Group in the United States and other countries.Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other countries, or both.Other company, product, or service names may be trademarks or service marks of others.

Information is provided "AS IS" without warranty of any kind.

The customer examples described are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics may vary by customer.

Information concerning non-IBM products was obtained from a supplier of these products, published announcement material, or other publicly available sources and does not constitute an endorsement of such products by IBM. Sources for non-IBM list prices and performance numbers are taken from publicly available information, including vendor announcements and vendor worldwide homepages. IBM has not tested these products and cannot confirm the accuracy of performance, capability, or any other claims related to non-IBM products. Questions on the capability of non-IBM products should be addressed to the supplier of those products.

All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only.

Some information addresses anticipated future capabilities. Such information is not intended as a definitive statement of a commitment to specific levels of performance, function or delivery schedules with respect to any future products. Such commitments are only made in IBM product announcements. The information is presented here to communicate IBM's current investment and development activities as a good faith effort to help with our customers' future planning.

Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon considerations such as the amount of multiprogramming in the user's job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve throughput or performance improvements equivalent to the ratios stated here.

Photographs shown may be engineering prototypes. Changes may be incorporated in production models.

Trademarks and Disclaimers

© 2013 ibm corporation networking fundamentals andy wright global field marketing, ibm system...

Documents