The Crown Jewels: Is Enterprise Data Ready for the Cloud?

Grab some coffee and enjoy the pre-show banter before the top of the hour!

The Briefing Room

The Crown Jewels: Is Enterprise Data Ready for the Cloud?

Twitter Tag: #briefr

The Briefing Room

Welcome

Host: Eric Kavanagh

[email protected] @eric_kavanagh


The Briefing Room

!   Reveal the essential characteristics of enterprise software, good and bad

!   Provide a forum for detailed analysis of today’s innovative technologies

!   Give vendors a chance to explain their product to savvy analysts

!   Allow audience members to pose serious questions... and get answers!

Mission


The Briefing Room

Topics

This Month: CLOUD

April: BIG DATA

May: DATABASE

2014 Editorial Calendar at www.insideanalysis.com/webcasts/the-briefing-room


The Briefing Room

Analyst: Robin Bloor

Robin Bloor is Chief Analyst at The Bloor Group

[email protected] @robinbloor


The Briefing Room

NuoDB

! NuoDB is a NewSQL distributed database solution

!   It is architected to scale elastically on the cloud

! NuoDB leverages a peer-to-peer distributed architecture, and it is ACID complaint and continuously available


The Briefing Room

Guest: Jim Starkey

Jim Starkey invented the NuoDB Emergent Architecture, and developed the initial implementation of the product. Jim’s career as an entrepreneur, architect, and innovator spans more than three decades of database history from the Datacomputer project on the fledgling ARPAnet to his most recent startup, NuoDB, Inc. Through the period, he has been responsible for many database innovations from the date data type to the BLOB to multi-version concurrency control (MVCC). Starkey has extensive experience in proprietary and open source software. Starkey joined Digital Equipment Corporation in 1975, where he created the Datatrieve family of products, the DEC Standard Relational Interface architecture, and the first of the Rdb products, Rdb/ELN. Starkey founded Interbase Software in 1984 and Netfrastructure, Inc. in 2000.

The Briefing Room Jim Starkey

March 25, 2014

Magic Quadrant 2013

NuoDB !   Next-generation distributed database !   Designed for cloud, datacenter, and on-premise

deployment !   Unique ability to deploy an active-active database

in multiple locations !   Deep database DNA in management team and

world-class investors !   Headquartered in Cambridge, MA

Dassault Systèmes Dassault Systèmes: !   2nd largest independent software

vendor (ISV) in Europe !   Leader in 3D design software, 3D

Digital Mock Up and Product Lifecycle Management (PLM) solutions

!   170,000 customers and 10M on-premise users

!   Customers include Boeing, Ford Motor Company, Guess apparel, NASA, Airbus, Fujitsu, Coca Cola and thousands of others

! NuoDB is an integral part of their cloud-based 3DEXPERIENCE strategy

!   Investor in NuoDB

“NuoDB delivers a lot of the features required to address the

market needs in terms of usages in the new world of experiences.”

“This investment demonstrates our strong interest and belief in

NuoDB’s strategy and technologies for next-generation

cloud based services.”

Dominique Florack, Senior Executive VP

Products-R&D Dassault Systèmes

Ø Rigid & Inflexible Ø Dedicated servers Ø Scale-‐up / No Scale-‐down

Ø Low u>liza>on Ø High Administrator/Applica>on ra>o

Ø Mul>ple single points of failure

Ø Maintenance down>me

Ø High capex Ø Single datacenter

Conven>onal Applica>ons

We need a distributed database system …

Ø Web Servers Scale-‐out ✓

Cloud-‐Style Applica>ons

Ø App Servers Scale-‐out ✓

Ø DBMS Servers don’t Scale-‐out✗

Ø Storage Servers Scale-‐out ✓

Can a RDBMS do this?

Time

TPS

(Without giving up SQL or ACID Transac>ons)

Jim Starkey

“Elas>cally Scalable Transac>ons represent

the biggest breakthrough in

database technology in 25 years”

Elastic Scale-out Multi-Tenancy

Continuous Availability

No-knobs Admin

Breakthrough Capabilities

•  NuoDB scales to over 100 server machines

•  Scalability is instant and elastic •  Scales-out and scales-in •  TPS numbers exceed 10m TPS on

$100k of hardware •  Also scales on AWS, GCE etc. Public

demo of 32 nodes with GOOGLE •  Now showing linear scalablity on

TPC-C type workloads (DBT-2) •  Scalability demonstrated with

heavier duty customer applications (eg Axway, Dassault Systémes)

•  Self-healing •  No single point of failure •  Fully distributed control •  Arbitrarily redundant •  Online backup •  Online schema evolution •  Rolling upgrades

•  HP Moonshot Launch – 45 Micro servers in a 4U rack mount box

•  NuoDB ran 72,000 databases on a single Moonshot box

•  Uses proprietary “Database Hibernation” and “Database Bursting” technologies

•  Zero admin UI •  Demo showed the potential of

“Software Defined Database” •  Moonshot is the foundation of

the HP relationship •  Active/Active •  ACID Semantics •  Transactional

Consistency •  N-Way Redundant •  Local User Latency •  Asynch WAN Comms

•  Auto-admin •  Rules-driven •  Auto-optimizing •  Auto-backup

Geo-Distribution

16

Multi-Tenancy

No-knobs Admin










Elastic Scale-out







Geo-Distribution

17



Elastic Scale-out

•  NuoDB scales to over 100 server machines •  Scalability is instant and elastic •  Scales-out and scales-in •  TPS numbers exceed 10m TPS on $100k of

hardware •  Also scales on AWS, GCE etc. Public demo

of 32 nodes with GOOGLE •  Now showing linear scalablity on TPC-C type

workloads (DBT-2) •  Scalability demonstrated with heavier duty

customer applications (eg Axway, Dassault Systémes)

Elastic Scale-out Multi-Tenancy


No-knobs Admin

















Geo-Distribution

18



Multi-Tenancy

No-knobs Admin










Elastic Scale-out







Geo-Distribution

19



•  Active/Active •  ACID Semantics •  Transactional Consistency •  N-Way Redundant •  Local User Latency •  Asynch WAN Comms

Geo-Distribution

Multi-Tenancy

No-knobs Admin










Elastic Scale-out







Geo-Distribution

20



No-knobs Admin



The Briefing Room

Perceptions & Questions

Analyst: Robin Bloor

The Quest of Many Database Engineers

True database distribution has always been a Holy Grail

HERE’S WHY…

What is a Database? A database is software that presides over a heap of data that:

IMPLEMENTS a data model

MANAGES multiple concurrent requests for data

IMPLEMENTS a security model

IS ACID compliant (?)

IS resilient

The Problem of Distribution

Databases Have to Distribute

Databases always scaled-out somewhat…

u  Usually it is best to scale up (on a single node) before scaling out

u  The first scale-out step is onto well-engineered cluster

u  Then onto a more loosely bound grid

u  At some point the scale-out sharding approach will run into bottlenecks, depending on workload

u  This will occur sooner with OLTP workloads

Approaches to Distribution…

PRIOR ATTEMPTS AT DISTRIBUTION:

Note that geo-distribution is

just distribution with bigger

latency issues

Simple replication

(master-slave)

Multi-master replication

(= peer replication)

If I understand it correctly,

NuoDB implements multi-master replication

u  You depict NuoDB as requiring zero admin. What parameters can the user set?

u  100 server nodes – what (roughly) is the latency penalty?

u  What is the latency penalty for geo-distribution, roughly speaking.

u  How well does NuoDB manage large query workloads?

u  Can you explain the recovery possibilities available with NuoDB?

u  What can you tell us about Dassault Systèmes’ use of NuoDB?

u  Why is NuoDB suited to cloud operation?


The Briefing Room


The Briefing Room

Upcoming Topics

www.insideanalysis.com

2014 Editorial Calendar at www.insideanalysis.com/webcasts/the-briefing-room

This Month: CLOUD

April: BIG DATA

May: DATABASE


The Briefing Room

THANK YOU for your

ATTENTION!

Images borrowed from the Internet: Slide 23: http://www.film-intel.com/2012/01/why-americans-like-monty-python-and.html

The Crown Jewels: Is Enterprise Data Ready for the Cloud?

Technology

twitter tag

briefrthe briefing roomanalyst

briefrthe briefing roomguest

briefrthe briefing roommission

briefrthe briefing roomnuodb

briefing roomjim starkeymarch

nuodb emergentarchitecture

peer distributed architecture