Copyright © 2014 Comcast Inc.
Jonathan Luste So<ware Tools Analyst, Comcast Cable
Video and Mobile Insights @ Comcast: Splunk for Today's AnalyKcs
Disclaimer
2
During the course of this presentaKon, we may make forward-‐looking statements regarding future events or the expected performance of the company. We cauKon you that such statements reflect our current expectaKons and
esKmates based on factors currently known to us and that actual events or results could differ materially. For important factors that may cause actual results to differ from those contained in our forward-‐looking statements,
please review our filings with the SEC. The forward-‐looking statements made in the this presentaKon are being made as of the Kme and date of its live presentaKon. If reviewed a<er its live presentaKon, this presentaKon may not contain current or accurate informaKon. We do not assume any obligaKon to update any forward-‐looking statements we may make. In addiKon, any informaKon about our roadmap outlines our general product direcKon and is subject to change at any Kme without noKce. It is for informaKonal purposes only, and shall not be incorporated into any contract or other commitment. Splunk undertakes no obligaKon either to develop the features or funcKonality described or to
include any such feature or funcKonality in a future release.
Agenda
! Who are you? Who am I? ! The Product(s) and the Problem(s) ! Ingredients for Success ! Grab Your Shovel ! Nuclear-‐Grade Enrichment ! Release the Hounds ! Show me Things over Stuff, Right Now! ! Lessons Learned, QuesKons Answered
3
Who Are You? Who Am I?
Who Are You?
5
Who Am I?
6
! Splunk Power User at Comcast Cable – Tech and Prod Development ! Group-‐imposed Ktles of Knowledge Manager and Data Czar ! Was Business / Technical Analyst for a while ! Did Vendor Management ! Degree in Business NOT (programming OR engineering)
Jonathan Luste
Who Am I?
7
The Product(s) and The Problem(s)
XFINITYTV.COM
9
XFINITY.COM
10
Visit the new XFINITY customer site to get the most out of your subscripKon. Watch TV online, check email, pay bill and upgrade Comcast subscripKon
XFINITY TV & TV Go Apps
11
XFINITY TV Remote Apps
12
XFINITY Home App
13
TradiKonal AnalyKcs Tools No real “One Size Fits All”
14
Inadequacies • Video and Audio health staKsKcs • Limited mobile applicaKon support • Cross data source & system
analysis • Deep-‐dive analyKcs • Inflexibility in real-‐Kme reporKng • Custom reports may require Pro
Services or be unavailable
Focus • Ecommerce • MarkeKng • CompeKKve intelligence • Single point transacKons • Mobile
– . . . browsing
Wheel of Awesome
15
Ingredients for Success
Splunk Tech Talk – Available Data to Use
17
! Sourcetypes: More than a hundred ! Hosts: Nearly a thousand ! Sources: Over 10 thousand ! Events: Greater than 50 Billion ! Benefit: All searchable together!
Example Product Eco System
18
Team 1
Team 2
Team 3
Grab Your Shovel
20
A Search for Stuff Super easy to look at one thing across a complex environ-‐ment
Nuclear-‐Grade Enrichment
ReporKng Problem
22
! The User Agent String reports on what each client is using as far as OperaKng System and browser
! Machine data can someKmes not mean much to non-‐technical teams
Screenshot here
Need to know OS versions that customers are using
SoluKon – Add Knowledge
23
! By tagging NT 6.1 as Windows 7, it turns data into informaKon
Screenshot here
UKlize Tags to enrich the WindowsOSVersion to something in English
Examples of Added Knowledge for Video
24
! Event types – Classify video events – Define and exclude edge cases
! Tags – Change error codes to error types
! Lookups – Add a video name from an video ID
! Macro – `Video01` searches just what you want
Release the Hounds
Example ProducKon Support Monitoring Problem X number errors per hour raises the flag for invesKgaKon, by error code
26
Failure Scenario 2
• Series finale of a popular show just aired and is available online, every hour for an enKre weekend over 300,000 people auempt to watch it
• Very low error rate, constantly alarming
Failure Scenario 1
• 188 Customers want to watch a movie during 11PM and 12AM, it is no longer available online but the URL exists
• 100% error rate, possibly missed
Example SoluKon – Monitor & Alert By creaKng a field from simple math, we can now look at error rates
27
Success Scenario 2
• Series finale of a popular show: • 25,000 every 5 min. If there are
over 500 errors (2%) let me know, otherwise ignore as it is traffic related
• 2% error rate or MORE ONLY! No false posiMves
Success Scenario 1
• 15 Customers want to watch a movie during 11:00PM and 11:05PM, it is no longer available online but the URL exists
• 100% error rate reported
Show me Things over Stuff, Right Now!
Product Insight & OperaKons Problems
29
Business OperaKons
• How many users are downloading videos?
• What app versions can be disconKnued from support?
• How many devices are registered? • Is traffic coming from live
streaming or on demand content?
Product
• How is my product being used? • What devices are being used? • What are the top models? • More iOS or Android? • Download vs. streaming?
SoluKon – Custom Dashboards
30
! View applicaKon uKlizaKon
! AND look at OS stats ! AND report on popular devices
! AND show performance stats
! AND count video stats ! AND AND AND AND
SoluKon – Custom Dashboards – More
31
! I need by OS Version!
SoluKon – Custom Dashboards – More
32
! I need by ApplicaKon Version!
SoluKon – Custom Dashboards – More
33
! I need by Device Make & Model!
Lessons Learned QuesKons Answered
Lessons Learned ! Splunk succeeds where tradiKonal web analyKcs tools fall short
– Changes are easily applied to reports – ReporKng potenKal in Splunk is limited by your own Kme and experience
! Splunk uses ALL of your data to be able to provide end-‐to-‐end visibility and provide answers to all types of business staff: – ProducKon Support – Product Design & Development – Business OperaKons – Quality Assurance
! Enriching your data and creaKng dashboards brings your product knowledge to life in the way people need it
35
QuesKons Answered
! Which app versions are customers running? ! What devices and operaKng systems should QA devote auenKon to? ! Performance, how many video errors compared to video starts? ! How should development prioriKze issues based on customer experience?
! Top video types; streaming versus downloaded staKsKcs
36
THANK YOU