presented

Analytics as a Service

OLIVER RATZESBERGER

Sr. Director Architecture & Operations

Cloud Computing, Analytics as a Service

eBay inc.

October 2008

eBay Inc. confidential2

eBay Inc Overview

As of December 31, 2007

• Founded in September of 1995, eBay is a global online marketplace where practically anyone can trade practically anything.

• eBay has a global presence in 39 markets, including the U.S.

• eBay has approximately 276 million registered users worldwide.

• Marketplaces net revenues totaled a record $1.5 billion in Q4-07, representing a year-over-year growth rate of 21 percent. With 46 percent from US operations and 54 percent from our International business.

• In Q4-07, gross merchandise volume (GMV), the total value of all successfully closed items on eBay's trading platforms, was $16.2 billion. (Total GMV for the full year 2007 was more than $59 billion.)

• eBay users worldwide trade more than $2,039 worth of goods on the site every second.


eBay Inc Overview (cont)

• There were 637 million new listings added to eBay worldwide in Q4-07. At any given time, there are approximately 113 million listings worldwide, and approximately 6.7 million listings are added per day. eBay users trade in more than 50,000 categories.

• At the end of Q4-07, eBay hosted approximately 532,000 stores worldwide, with approximately 46 percent of stores hosted on eBay's international sites.

• eBay members worldwide have left more than 6 billion feedback comments for one another regarding their eBay transactions.

• The most expensive item sold on eBay to date is a private business jet for $4.9 million.


Velocity of trading

On an average day on eBay…

A Diamond Ring is sold every two minutes


Velocity of trading


More than 3 Watches are sold every minute


Velocity of trading


5 Women’s handbags are sold every minute


Velocity of trading


Over 3600 MP3 players are sold


Velocity of trading


A makeup product sold every 2 minutes


Velocity of trading


4,827 fragrance products sold per day


Velocity of trading


A hair product sold every second


Velocity of trading


Over 300 stamps are sold every hour


Velocity of trading


An automobile is sold every minute


eBay Analytics Technology Highlights

>50 TB/day of new, incremental data

>50 PB/day >50^10 new records/day

Millions of queries/day

>5000 business users & analysts

>50k chains of logic

24x7x36599.9+% Availability

turning over a TB every5 secondsActive/Active

Near-Real-time

>100k data elements

Always online

Processed


eBay Analytics Core

Relational Data

XML, name/value, raw

Teradata Teradata

2.2PB 6.6PB

Primary Secondary

Phoenix, AZ Sacramento, CA

Linux Linux

Solaris

2.2PB2.5PBMPP

MPP

MPP/HPC/GridMPP/HPC/Grid

Sun Fire 4xxx

Relational Data

Local Interconnect

MicroStrategy

Wide AreaInterconnect1000 miles

Solaris

Business Objects

Unica SOA/DAL

Ab Initio Informatica UC4 SOA

Data Access

Data Integration

Crystal SAS SQL

Sun Fire 4xxx

Local Interconnect

Golden Gate BES

MAX

MAX


Analytics DNA

Embedded in our daily life

Bottoms-up & Tops-down

Think and Live Analytics

Always

But know when do avoid Analysis Paralysis!


Types of Analytics at eBay

Basically measure anything possible - A few examples:


Key Performance Indicators

Align individual and

departmental performance objectives

with corporate

goals


KPI Example: Technology Operations

Parallel Efficiency – (simplified) The effectiveness of distributing large amounts of workload over pools and grids of servers.

100% is GOOD Less than 70% is BAD

10,000 Server running at an average PE of 50%

Established through Analytics of Operations Data – Minute by minute utilization metrics of entire infrastructure

Raising PE from 50% to 80% equals Millions in OpEx savings


KPI Example: Technology Operations (cont)

• Individual process rollup• Grid level Parallel Efficiency at 99.9%


Design for the Unknown

>85% of eBay analytical workload is NEW & Unknown

Exploration is the core of an analytical company

The metrics you know are ‘cheap’

The metrics you don’t know are expensive but also high in potential ROI

Design can’t be static or dependent on specific questions or dimensions


Proliferation of Analytics

Decentralized Analytics

Shortened Time to Market Requirements

Adhoc Exploration

Departmental Data

Prototyping - Can’t wait for EDW

”We Need Data Marts!”


Proliferation of Analytics

Hub and Spoke Architecture ‘The Solve’


Data Mart Dilemma

Total Cost of Ownership (TCO)Fully loaded cost staggering $500k++

Biggest drivers are

Maintaining separate databases

weekly/daily/hourly data transfers

Data inconsistencies Data redundancyIncreased complexity Loss of lineage over time

A Data Mart cannot be ‘cheap’ enough to justify its existence


Agile Analytics needs Analytics as a Service

Massive scale Analytical Utility Computing

Bring your data - Perform your Analytics

From Simple Web based data upload

...to fully private Utility access

Combine custom data and code with ALL existing data



From simple web based table upload



...to fully private utility access

We call them PET (Prototyping Environment = Sandbox)

More than 75 active right now

In most cases they are small (<500GB)

since all the main data is already in the EDW

They are free to the business units


Analytics as a Service - Benefits

Improved Time To Market - Days/Weeks vs Months

Enable the business to do agile prototyping

Enable the users to “Fail Fast” - Make it easy

to try out new ideas

Eliminate stray Data Marts


Resource Budgeting for Business Units

Aid in Enterprise Capacity Planning

Enable Agile Analytics as a Service29

Resource Allocation Model (RAM)

Activity Based Costing Model


Questions

?


Find out more....

On our recently launched technology blog:

presented

Documents

Transcript of presented