Use More, Pay Less - Delivering Increased Storage Efficiency without Compromising on SLAs

Use More, Pay Less – Delivering

Increased Storage Efficiency

without Compromising on SLAs

Adrian Groeneveld

Marketing Director, EMEA

Market Dynamics

Economic meltdown

Virtualisation is mainstream

Cloud computing on everyone’s discussion agenda

Unified everything

Companies forced to adopt a new approach

Storage at the centre of everything

The Storage Issues

Silo’d or fragmented storage through project driven

acquisition

Poor utilisation

Highly inefficient due to power, space and resource

overhead

Legacy storage scales poorly, creating bottlenecks

Sharing storage between applications creates contention

and missed SLAs

Building a Shared Infrastructure

Private Cloud

– Typically an internal shared compute environment to support

business units/divisions and internal applications

Public Cloud

– Typically an external organisation using their own infrastructure

to deliver services to multiple clients e.g. SalesForce.com

Cloud Benefits

– Efficiency of resource

– Flexible shared environment

– Simple resource allocation

– Simplified management

– Project based charge back

Obstacles to Successful Shared

Storage Infrastructure

How unified is unified?

Applications don’t like sharing

Issues around scalability

Who’s using what

Securing the Infrastructure

Allocation of resources

Resource prioritisation

“Pillar delivers an innovative next generation

data storage solution enabled by patented

Quality of Service technology and built on a

highly scalable modular architecture to

automatically regulate and manage

fulfillment of incoming application data

requests based on business importance.”

One Key Enabling Technology:

Pillar Quality of Service

Enabling Cloud Storage

Enabling the Virtual Datacentre

Enabling Faster Applications

Enabling More Usage with Less Costs

9

How Do We Do It?

One Platform: Axiom 600

Key Value: IO Prioritization

Virtual Server

Virtual Machine 2

Virtual Machine 1

Virtual Machine 3

1 3

9

4 5

8

10

7

2 6

Medium Priority Queue

high medium low

High Priority Queue

Low Priority Queue

Core Technology: QoS

Architecture Components

Storage Controllers - Slammer • Active-Active SAN or NAS control units

• 48GB Cache

• (8) GigE ports or (4) 4 or 8Gbps FC ports

• Complete High Availability

• Scales up to four Slammers (eight control units)

Storage Enclosure - Brick • Dual RAID controllers

• Embedded RAID-5 and RAID-10

• SATA, Fibre Channel and SSD drives

• Built in hot spare

• Hot-swappable drives and RAID Controllers

Management Controller - Pilot • Dual control units

• Active/Passive - High Availability

• (2) Hard drives for Config and Logging

• (6) 10/100/1000 Base-T Ethernet

One Platform: Axiom 600

4x SAN and/or NAS Slammers – SAN (FC, iSCSI)

4-16 4Gbps and/or 8Gbps FC ports

4-16 1Gbps Ethernet ports

– NAS (NFS or CIFS)

8-32 1Gbps Ethernet ports

– 48-192 GB of Cache

64x Bricks – 1.66 PB of Capacity

– Storage Classes

SATA: 500GB, 1TB and 2TB

FC: 300GB and 450GB

SSD: 50GB and 200GB

Guarantees – Performance guarantee with 80%

Utilization

Pillar Axiom Modular Scaling

Pilot

Slammer

Slammer

Bricks

True Performance Scalability: Proof

Differentiating Technologies

Quality of Service (QoS) – Dynamic CPU, Cache, and Capacity Management

Unified Storage – Native SAN & Native NAS shared storage pool

Distributed RAID – Linear Scaling, Performance under fault, fast rebuild

Policy based Management & Application Aware Profiles – Storage services linked to application priority

Modular architecture – Scale with highest utilization rates in enterprise storage

Quality of Service

Problem: Traditional arrays still use a FIFO (First in First

Out) I/O queuing model

– No regard for the business importance of an application – first

come first served

Solution: Prioritize I/O’s by the business value of an

application

– Multiple I/O queues: Important I/O’s executed first regardless of

received order

Benefit: Aligns performance with business needs vs.

who got there first

How we do it: QoS Control Points

Storage Domains – Extending QoS

Isolates specific bricks within the storage pool

Usage

– Database Environments - Create separate domains

for the logs and the data files for data protection and

comfort

– Hosting – Allocate specific bricks to specific

customers within the overall Axiom pool

– Private cloud – allow departmental charging for bricks

– Overcome project based storage acquisition

– Brick based data migration for brick upgrades

Extending QoS for Share Storage

Environments

Slammer acts as a type of Cloud Controller

Each domain has its own Bricks, Storage Classes and

RAID controllers

Storage Domains act as separate application containers

A separate storage domain can be created for every

Cloud user (application owner or external cloud

customer in a hosting example)

Users can acquire additional bricks to be added to their

own storage domain within the Cloud

Every Cloud user still has the benefits of Quality of

Service and storage classes

A Customer Example

Canterbury City Council

“Our guys aren’t spending all their

time fire fighting...routine is

back to routine!”

Timo Bayford, Computer Services Manager

Delivering Shared Storage Services

700 full-time staff

25 sites to support

70 business units

Massive range of services

– Council tax calculation and collection

– Theatre and festival ticketing

– Car parks

– Benefits payments

– Many more

The Challenges Faced

Existing system was having constant performance

issues

Mixed drive types was creating contention

Needed a system to cope with multiple applications that

could cope with growth

Were faced with constant complaints from the business

units

Revenue generating systems were getting impacted

Need to be able to deliver shared services to

neighbouring councils

The Solution Deployed

Systems Deployed

– Axiom 600, dual Slammer

– Both with native SAN and NAS

– Fibre Channel and SATA drives

Applications Serviced – Exchange

– SQL

– Oracle

– ESX Server

– Data archive and back-up

– Finance, Web and Billing

– Other

The Benefits

Instant provisioning and dynamic change based on application

demand with Pillar’s Quality of Service

Supporting multiple applications on physical and virtual servers with

predictable and consistent performance managed with QoS

Virtually eliminating complaints from users due to poor response

from application requests

Ability to scale and deliver shared services architecture for reduced

TCO amongst a number of councils

Consolidation of all applications onto a single, unified storage

platform, without impacting the business

Building Your Shared Infrastructure

And Finally

WIN AN iPad

Come to Stand 559

Answer 3 simple

questions

This could be yours

Thank You Questions?