Jon Hudson - The Evolving Role of the Network Engineer

@2015 BROCADE COMMUNICATIONS SYSTEMS, INC. VIOLATORS WILL BE HUNTED BY DRONES

Jon HudsonPrincipal Engineer & FuturistOffice of the CTO, Brocade

The Evolving Role of the Network Engineer


…Into the Rabbit Hole


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. ADX, AnyIO, Brocade, Brocade Assurance, the B-wing symbol, DCX, Fabric OS, HyperEdge, ICX, MLX, MyBrocade, OpenScript, VCS, VDX, and Vyatta are registered trademarks, and The Effortless Network and The On-Demand Data Center 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 others.

Legalish Disclaimer

@2015 BROCADE COMMUNICATIONS SYSTEMS, INC. VIOLATORS WILL BE HUNTED BY DRONES 3

This presentation, with prepared remarks, includes forward-looking statements regarding Brocade’s financial results, goals, plans, assumptions, strategy and business outlook, revenue, tax rate, cash, stock buybacks, OEM inventory, prospects, as well as U.S. Federal spending and order flow, and IT spending, which are only predictions and involve risks and uncertainties such that actual results may vary significantly. These and other risks are set forth in more detail in our Form 10-K for the fiscal year ended October 26, 2013 and our Form 10-Q for the fiscal quarter ended August 2, 2015. These forward-looking statements reflect beliefs, assumptions, outlook, estimates, and predictions as of today, and Brocade expressly assumes no obligation to update any such forward-looking statements whether as the result of new developments or otherwise.In addition, this presentation includes various third-party estimates regarding market share and other measures, which do not necessarily reflect the views of Brocade. Further, Brocade does not guarantee the accuracy or reliability of any such information or forecast.Certain financial information is presented on a non-GAAP basis. Management believes that non-GAAP financial measures used in this presentation allow stakeholders to gain a better understanding of Brocade’s comparative operating performance both from period to period, and to its competitors’ operating results. Management also believes these non-GAAP financial measures help indicate Brocade’s baseline performance before gains, losses or charges that are considered by management to be outside of ongoing operating results. Accordingly, management uses these non-GAAP financial measures for planning and forecasting of future periods and in making decisions regarding operations performance and the allocation of resources. The most directly comparable GAAP information and a reconciliation between the non-GAAP and GAAP figures are provided in the accompanying press release, which has been furnished to the SEC on Form 8-K and posted on Brocade’s website, and is included in the appendix in this presentation.

Legal DisclosurePlease see risk factors on Form 10-Q and Form 10-K filed with the SEC


BROCADE > Futurist & Principal Engineer | Office of the CTOIETF > Co-Chair TRILL WG

> Routing Directorate Co-Coordinator > Routing Area Encapsulation Design Team Member> Architect/Design Group member for the NVo3 WG

ISO/IEC JTC 1 > JTC1BDSG - Study Group on Big Data (completed 1/2015)

INCITS > Ad Hoc on Big Data (completed 1/2015)

Jon Hudson

Recent Evidence of My Existence….

JON.HUDSON@[BROCADE]||[GMAIL].COMSMS: 1.650.766.7662@THE_SOLUTIONEER @2015 BROCADE COMMUNICATIONS SYSTEMS, INC. VIOLATORS WILL BE HUNTED BY DRONES 5


JON.HUDSON@[BROCADE]||[GMAIL].COMSMS: 1.650.766.7662@THE_SOLUTIONEER 6

SERVER L2-7 NETWORK GEAR ROUTING & POLICY ORCHESTRATION

Software Defined DatacenterOn Demand DatacenterRight Click, Datacenter

DAYS DAYS DAYS DAYS

X86 VirtualizationX86 Virtualization

MINS

NfVNFV

MINS

PROGRAMMATICOpenFlow

MINS

API AUTOMATIONOpenstack

MINS

OPEN Daylight



STAGE 0 HUMAN ALERT NETWORKSTAGE 1 LOG EVERYTHING, DON’T ASK WHY STAGE 2 ANALYTICS & SOFTWARE REMEDIATIONSTAGE 3 MACHINE LEARNINGSTAGE 4 FIND NEW CAREER

STAGE 0PATIENT 0 DIES…AND PROBABLY A FEW MORE…

BUILD A DATACENTER

DEPLOY AN APPLICATION

USER FINDS AN ISSUE

IT GUY DEBUGS

LOG EVENT & USER EVENT CORROLATION

WAIT FOR CONFIRMATION



HUMANS ARE BAD MONITORING TOOLS

STAGE 0


BUILD A DATACENTER


USER STILL FINDS AN ISSUE

IT GUY DEBUGS



SETUP BASIC KNOWN TRIGGERS

LOG EVERYTHING!

STAGE 1PATIENT 0 STILL DIES…BUT HOPEFULLY NOT AS MANY


HOW DO YOU MONITOR THE KNOWN UNKNOWNS?

A SCSI BACKPLANE ISSUE ON EMC SYMMETRIX STORAGE ARRAY DISCOVERED SIX MONTHS PRIOR TO ARRAY BASED ALERTING BY TRACKING “BURSTABLE EFFICIENCY” OVER TWELVE MONTHS

DB LOSS TRACKED OVER LONG PERIODS TO INDICATE LIKELY TRANCEIVER DEATH

2 YEARS OF DATA NEEDED TO CATCH ALL HUMAN CYCLES

STAGE 1


BUILD A DATACENTER



IT GUY DEBUGS




STAGE 2—ANALYTICSPATIENT 0 HAS A TREATMENT PLAN READY ON STANDBY

IP Analytics Pack for vRealize OperationsVirtual network resilience through physical network visibility

VIRTUAL NETWORK RESILIENCE THROUGH PHYSICAL NETWORK VISIBILITY

Intelligent Network OperationsDelivers comprehensive VCS operational data to support software-defined data center analytics

VM

Unified Virtual & Physical NetworkingProvides insights and visibility into the physical network for better application and VM performance

Simplified Policy AutomationPredefined alerts and recommended actions simplify monitoring; health, risk, and efficiency metrics add context



STAGE 2PATIENT 0 HAS A TREATMENT PLAN READY ON STANDBY


BUILD A DATACENTER



IT GUY DEBUGS


RINSE & REPEAT


*Want to know more about Machine Learning? Please Check out David Meyer’s ML Presentationhttp://www.1-4-5.net/~dmm/talks/2015/MWC/mwc2015.pptx

STAGE 3—MACHINE LEARNING*BRINGS YOU NEW QUESTIONS

Google PUE Optimization Application1

• Straightforward application of ANN/supervised learning– Lots more happening at Google (and FB, Baidu, NFLX, MSFT,AMZN,…)• http://research.google.com/pubs/ArtificialIntelligenceandMachineLearning.html

• Use case: Predicting Power Usage Effectiveness (PUE) – Basically: They developed a neural network framework that learns from

operational data and models plant performance– The model is able to predict PUE2 within a range of 0.004 +

0.005 , or 0.4% error for a PUE of 1.1• “A simplified version of what the

models do: take a bunch of data, find the hidden interactions, then provide recommendations that optimize for energy efficiency.”

– http://googleblog.blogspot.com/2014/05/better-data-centers-through-machine.html

1HTTPS://DOCS.GOOGLE.COM/A/GOOGLE.COM/VIEWER?URL=WWW.GOOGLE.COM/ABOUT/DATACENTERS/EFFICIENCY/INTERNAL/ASSETS/MACHINE-LEARNING-APPLICATIONSFOR-DATACENTER-OPTIMIZATION-FINALV2.PDF

2 HTTP://EN.WIKIPEDIA.ORG/WIKI/POWER_USAGE_EFFECTIVENESS @2015 BROCADE COMMUNICATIONS SYSTEMS, INC. VIOLATORS WILL BE HUNTED BY DRONES 17

Google Use Case: FeaturesNumber of features relatively small (n = 19)


Google Use Case: Details

• Neural Network– 5 hidden layers– 50 nodes per hidden layer– 0.001 as the regularization parameter (λ)

• Training Dataset– 19 normalized input parameters (features) per

normalized output variable (the DC PUE)• Data normalized into the range [-1,-1]

– 184,435 time samples at 5 minute resolution• O(2) years of data

– 70% for training, 30% for cross validation


Google Use Case: Sensitivity Analysis

• After the model is trained, one can look at effect of individual parameters by varying one while holding the others constant

The relationship between PUE and the number of chillers running is nonlinear because chiller efficiency decreases

exponentially with reduced load.


Google: Outside Air Enthalpy Has Largest Impact on PUE

Relationship between PUEandoutside air enthalpy, or total energy content of the ambient air. As the air enthalpy increases, the number of cooling towers, Supplemental chillers, and associated loading rises as well, producing a nonlinear effect on the DC overhead. Note that enthalpy is a more compre-hensive measure of outdoor weather conditions than the wet bulb temperature alone since it includes the moisture content and specific heat of ambient air.



STAGES OF ENLIGHTENMENT

BUILD A DATACENTER


USER FINDS AN ISSUE

IT GUY DEBUGS



STAGE 0 HUMAN ALERT NETWORK



BUILD A DATACENTER



IT GUY DEBUGS




LOG EVERYTHING!

STAGE 0 HUMAN ALERT NETWORKSTAGE 1 LOG EVERYTHING,

DON’T ASK WHY



BUILD A DATACENTER



IT GUY DEBUGS





DON’T ASK WHYSTAGE 2 ANALYTICS



DON’T ASK WHYSTAGE 2 ANALYTICS STAGE 3 MACHINE LEARNING


BUILD A DATACENTER



IT GUY DEBUGS


RINSE &REPEAT



EVIL SOFTWARE COMING TO TAKE YOUR JOBS

STAGE 4: FIND NEW CAREER

Tale of the Tape2000 vs. 2015

TALE%OF%THE%TAPE%2015%

1630lbs ----- Weight ------- 59lbs

4x240v ------- Power ------- 2x120v

32100 ---- BTU ------ 2697

Yes --- 10G/40G Line-Rate --- Yes

Yes --- 100G Line-Rate --- Yes*

YES --- BGP/MPLS --- YES

No --- SW/HW Separation --- Yes

Custom -------- Type ---- Merchant

IOS XR -------------- OS ------- You Choose

> $1M --- Cost --- ~$5k

> $200k --- 1yr Support --- $2k

Yes ----Hot Swap HW ---- No

Yes --- COD --- No

Yes --- HW Partitions --- No

%

TALE%OF%THE%TAPE%2000%

1200lbs ---- Weight ------- 59lbs

4x240v ------- Power ------- 2x120v

32100 --- BTU ----- 2697

64bit --------- Arch ---------- 32bit

24 --------- CPUs ------------- 2

48 ---------- Threads ------------ 2

384GB ---- Memory ------- 32G

Custom -------- Type ---- Merchant

Solaris --- OS --- Linux

~$250k ----- Cost ------ ~$2.5k

$60k --- 1yr Support --- $2k

Yes ----Hot Swap HW ---- No

Yes --- COD --- No

Yes --- HW Partitions --- No

%%%

2000 2015


THIS IS ALSO ABOUTMULTITENANCY


Next Gen Data CentersDirect networking and free range VMs

x20






ProgrammabilityPython scripting


To use Python interpreter: NOSCLI “python” from NOSCLI shell.

Create Python script; copy script to switch and execute

sw0# copy scp://releaseuser:[email protected]//users/home32/badaniya/proj/scorpius/scripts/create_po.py flash://create_po.py

sw0# python create_po.py !Command: show running-config int vlan!Time: Tue Jan 6 08:32:33 2015

interface Vlan 1!

Script can be loaded on principal switch and automate VCS Fabric

34

ProgrammabilityVDX Puppet Integration

Node1

Node2 Netconf

Netconf

Puppet Master

VCS1

Catalog

Catalog

catalog

VCS2

Catalog

Catalog

catalog

• nos_netdev_device: Device Configuration• nos_netdev_interface: L2 Interface configuration• nos_netdev_vlan: VLAN configuration• nos_netdev_l2_interface: VLAN to interface assignments

NOS Provider List



Real-Time Routing with Reality Don’t guess, know!


Analysis Engine

37



Thank you

Government & Nonprofit

Jon Hudson - The Evolving Role of the Network Engineer