shunra university 1 intro to network virtualization

© Copyright 2013 Shunra Software, LTD. Information contained herein is subject to change without notice.

Shunra University Session 1

Introduction to Network Virtualization


Let’s Begin with a Bit of Theory

http://www.google.co.uk/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&docid=J6LVx4NwwnSyYM&tbnid=ig6L31ZL5Z6zBM:&ved=0CAUQjRw&url=http://joan-druett.blogspot.com/2010/12/kissing-and-albert-einstein.html&ei=rsCtUYzKDYKI9gSBsoDYBA&bvm=bv.47244034,d.eWU&psig=AFQjCNEAbE3fBx-AGmmqDx72gk4jPJrzpQ&ust=1370427909092776

www.shunra.com Network Virtualization for Software Testing

The Network Impact

3

Consider 2 Users - 1 in NY, 1 in Tokyo–Both access the same application in the NY Data Center–Will they all experience the application the same way?– If not, Why not?

TokyoLondonNew York

New YorkData Center

WAN


Network Impairments• Conditions which inhibit the flow of

data across a network• Key Impairments are:

– Network Delay– Jitter (Delay variation, inter-packet gap

variation)– Packet Loss (inverse of delivery rate)– Bandwidth Constraints– Bandwidth Utilization

4


The Mobile Network Impact

5

Consider 2 Mobile Users in NYC, 1 on WiFi one on 3G–Both access the same mobile application in the NY Data Center–Will they all experience the application the same way?– If not, Why not?


Mobile Network Impairments

• Mobile Networks Experience Extreme Conditions of the Same Net Impairments– More Network Delay (Last Mile + Backhaul,

RRC latency)– More Jitter (Delay variation, inter-packet gap

variation)– More Packet Loss (inverse of delivery rate)– More Bandwidth Constraints– More Bandwidth Utilization

6


Why is There Delay on a Terrestrial Network?

• Network Delay– Propagation Delay– Processing / Insertion /

Serialization Delay– Queueing Delay

A Bt (msec) = d / (2/3)c

Signal

d

Packet PacketPacket Packet PacketBUSY

Lost Packet

7


New York

London

Frankfurt

Tokyo

Sydney

Signal Speed – MPLS WAN

• Distance

• TheoreticalMinimum Signal Time(One Way)

• Speed of Light = 299,792.458 km/s

New York London Frankfu

rt Tokyo Sydney

New York

5,585km

6,219km

10,871km

15,998km

London 639km 9,585km

16,991km

Frankfurt

9,357km

16,480km

Tokyo 7,793km

New York

London Frankfurt

Tokyo Sydney

New York 28ms 31ms 54ms 80ms

London 3ms 48ms 85ms

Frankfurt 47ms 82ms

Tokyo 39msT = D / (2/3)c

8


Internet Consumer Latency

Broadband Latency (FCC)

ConnectionType

RTT

Fiber 17ms

Cable 28ms

DSL 44ms

http://www.fcc.gov/measuring-broadband-america

“Fiber-to-the-home services provided 17 ms round-trip latency on average, while cable-based services averaged 28 ms, and DSL-based services averaged 44 ms.”

9


Mobile Latency – Last Mile + Backhaul

10

For more on this topic checkout:Ilya GrigorickMobile Performance from the Radio Up session at Google IO


Mobile Latency – 1st. Data point

Ofcom study, UK Mobile Networks, May ‘11

http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/bbspeeds2010/Mobile_BB_performance.pdf

Static Tests:

User Panel:

“Latency… was an average of 117 milliseconds. Nearly two thirds of tests took more than 100 milliseconds”

“Average latency was 192 milliseconds and around a third of tests delivered latency of higher than 200 milliseconds”

11


Mobile Latency – 2nd. Data point

Cisco Cloud Index, 2010-2015

“Global average fixed latency is 125 and average mobile latency is 290… Global mobile consumer average latency is 307.3 ms”

12


Mobile Latency – Summary of Averages

13


Packet Loss

Packets get dropped on an end to end link due to the following:–Network congestion causes queues to saturate

–Network congestion triggers RED

– Interference on wireless networks

–Hardware errors

–Firewall Policies

Packet loss is the inverse of delivery rate–100% - packet loss rate = delivery rate

Random Early DetectionAverage queue monitored size & packets dropped based on statistical probabilities. If buffer is almost empty, all incoming packets are accepted. As queue grows, probability of dropping incoming packet grows. When buffer is full, probability reaches 1 and all incoming packets are dropped.

14


Packet LossPacket Loss, Akamai/MIT Research

Tested 149 Million connections in 51 countries– Compared PL for mobile networks from non-mobile networks

Most connections have no packet loss (Median PL is 0%).– 75th percentile usually zero too

Average connection packet loss: 2.5%– Fixed network average is 2.2%

Implies connections are usually fine, but sometimes terrible– The average PL for the non-zero half is 5%

Suggests “small” connections have higher Packet Loss– HTTP Connections are often small

http://pam2012.ftw.at/papers/PAM2012paper6.pdf

15




Jitter

Jitter = Variability over time of latency across a networkPDV – Packet Delay VariationConstant Latency = Zero Jitter

Jitter

No Jitter

Jitter-Sensitive Applications:• Video• VoIP

16


Bandwidth

17

Bandwidth = Network CapacityThink of number of lanes on the highwayConstraints the following:–number of user sessions * application throughput of each session

LAN Capacity

128Kbps - 5 Mbps

WAN Capacity (examples)

100 Mbps or

1 Gbps

T3 - 45 Mbps

T1 – 1.5 Mbps

Mobile Capacity (examples)


Mobile Bandwidth in Theory

18

Protocol Max Downlink Max Uplink

GPRS 56-115 Kbps

CDMA200 (1X) 153.6 Kbps

Edge 236.8 Kbps

3G Bar 2 Mbps stationary client384 Kbps moving client

EV-DO 2.5-4.9 Mbps 0.15-1.8 Mbps

EDGE Evolution 1.6 Mbps 0.5 Mbps

HSDPA 2 Mbps 200 Kbps

HSPA+ 14.4 Mbps

4G Bar 1 Gbps stationary client100 Mbps moving client

HSPA+ Release 11 21-672 Mbps 5.8-168 Mbps

LTE 100-300 Mbps 50-75 Mbps

WiMax 37-365 Mbps 17-376 Mbps


Mobile Bandwidth

19

You only get those speeds it if you’re this guy…


Mobile Bandwidth

20

PCWorld/Novarum3G–Download: 0.59 – 3.84 Mbps–Upload: 0.56 – 1.44 Mbps

4G–Download: 2.81 – 9.12 Mbps–Upload: 0.97 – 5.86 Mbps

Date: Apr, 2012

Source (PCWorld): http://bit.ly/JX7MQw

http://bit.ly/JX7MQw

http://bit.ly/JX7MQw


Test Time!

How much impact does the difference in Bandwidth & Latency have on web performance?

Tested Sites: Alexa US Top 100Details:–Increase Latency, Download & Upload Bandwidth separately

–Set no limits on other two variables

–Run each test 3 times, use median result


Download BW Impact on Load Time

DSL

Cable FIBER

Poor Cell Conn

Avg Cell Conn


Upload BW Impact on Load Time

DSL

Cable FIBER

Poor Cell Conn Avg Cell Conn


Latency Impact on Load Time

DSL

CableFIBER

Poor Cell Conn

Avg Cell Conn


Bandwidth/Latency Difference – Conclusions for Web Performance

Upload difference doesn’t matter much until <200 KBPS

Download difference DOES matterLatency difference matters the most


The Impact of Performance Optimization

26

Consider the 3G Mobile User in NYC, Running 2 Mobile Banking apps–Both apps run from the same handset at the same time–Will both apps perform the same way?– If not, Why not?–Hold that thought until we get to the analysis session


Knowing all that…

How are Organizations Testing Today?


Current Performance Testing Vs. Production

DevicesSite Type

Branch Office

Residential

Mobile (Anywhere)

NetworkData CenterFirewall

Road Warrior

28

Mobile Salesforce

WANVPN

DSL/CableWiFi

4G 3G 2.5GEtc.X X

XX

XXAutomated

Testing

Manual Testing

Load Testing

www.shunra.com Network virtualization for software testing.

MyAT&T Mobile App – Prod Vs. Test

WiFi Based Lab

Mobile Customers

3GLTEEdgeWiFi

AT&T Data Center

myAT&T – Dev/Test

AT&T Cloud Partner Services

WiFi based mobile tester

01-Launch_App

02-Login

03-View_Bill

04_Download_PDF

05-Usage

06-My_Plan

07-View_Line

08-Logout

02468101214

1.6

11.6

5.36.8

4.4 4.1

9.2

1.3

MyAT&T iPhone Application - Common User Actions

Wifi

http://www.google.com/url?sa=i&source=images&cd=&cad=rja&docid=SC3lgf-YUV4CaM&tbnid=oTzemGMhVQVIgM:&ved=0CAgQjRwwAA&url=http://warren3300.blogspot.com/2010/07/cell-towers-in-blowing-rock-nc.html&ei=gxB4UurXD4LMsQS8nIDICA&psig=AFQjCNHtWjWJxns-cIrJP8yDkcbisBnPyw&ust=1383686659305052


http://www.google.com/url?sa=i&source=images&cd=&cad=rja&docid=RwUXQUlmgbuZmM&tbnid=1UoMD4SKfR4h0M:&ved=0CAgQjRwwADhX&url=http://www.idownloadblog.com/2012/09/12/iphone-5-press-shots/&ei=lxJ4UtzAMKnOsASFlIHIBQ&psig=AFQjCNFX7qyzFSO2zJ2UJ5_4b1VilpMjvw&ust=1383687191840128


MyAT&T Mobile App – Prod Vs. Test

Mobile Customers

3GLTEEdgeWiFi

ATT Data Center

HP - Virtual Users

myAT&T – Dev/Test


Hi Speed Lab Network

http://www.google.com/url?sa=i&source=images&cd=&cad=rja&docid=IRolbSvd637F1M&tbnid=PgqMJtveq5DDYM:&ved=0CAgQjRwwAA&url=http://www.qaelearn.com/wordpress/?page_id=229&ei=hg94UpWYGfO_sQTG9YDwBQ&psig=AFQjCNGcbbB0oZHhVWgbdThGTHWlO4qHPg&ust=1383686406452574




Mobile End User Experience Testing with Shunra

Mobile Customers

3GLTEEdgeWiFi

ATT Data Center

Shunra Mobile Virtualization

myATT – Dev/TestShunra mobile tester


SNV-M



http://www.google.com/url?sa=i&source=images&cd=&cad=rja&docid=RwUXQUlmgbuZmM&tbnid=1UoMD4SKfR4h0M:&ved=0CAgQjRwwADhX&url=http://www.idownloadblog.com/2012/09/12/iphone-5-press-shots/&ei=lxJ4UtzAMKnOsASFlIHIBQ&psig=AFQjCNFX7qyzFSO2zJ2UJ5_4b1VilpMjvw&ust=1383687191840128


Mobile End User Experience Testing with Shunra

Mobile Customers

3GLTEEdgeWiFi

ATT Data Center

Shunra Mobile Virtualization

myATT – Dev/TestMobile Labs Integration


SNV-M



shunra university 1 intro to network virtualization

Technology