Achieve More. SanDisk Confidential 1 CMDDS Template Input Guidelines SanDisk EMS 19 June 2013.
Storage Solutions for Healthcare and Life Sciences · 9/6/2016 2 Long Live Data™ HGST and WD are...
Transcript of Storage Solutions for Healthcare and Life Sciences · 9/6/2016 2 Long Live Data™ HGST and WD are...
©2016 HGST, Inc. All rights reserved.
Storage Solutions for Healthcare and Life SciencesDavid Hiatt, Market Development - Healthcare and Life Sciences
August 23, 2016
9/6/2016 2
Long Live Data™
HGST and WD are independent subsidiaries of Western Digital Corporation
Sandisk is now integrated into both subsidiaries
Rich Heritage
Powerful Platform
Company Formation
©2015 HGST, Inc. All rights reserved. Confidential
9/6/2016
Long Live Data™
3
Executing on Our Strategic PrioritiesRecent Acquisitions
Expanding our capabilities and investing for ongoing growth
9/6/2016
Long Live Data™
4©2016 HGST, Inc. All rights reserved.
SOFTWARE
ClusteringManagement
Storage Analytics
STORAGEACTIVE ARCHIVE
Data AnalyticsBackup and ArchiveCloud Infrastructure
CAPACITY SCALE HDD
Backup – Object Storage
CAPACITY HDD
Cloud Storage – Hyperscale
SAS SSD
Online Transactions – HPC
PERFORMANCE HDD
Processing – Content Serving
NVMe
Databases &Analytics
Data Center Product Portfolio
HGST innovates at every level
of the storage stack – from the
fastest solid-state drives to the
densest storage systems
available.
9/6/2016
Long Live Data™
5
Our Connection to Life Sciences
©2015 HGST, Inc. All rights reserved. Confidential
Private testing centers– Partnering with large biopharmas
– 23 and Me / Pfizer depression study
New business models– Data brokers
Precision medicine– Enabled by digital pathology and genomic
sequencing
– Connecting systems for single view of patient
Interest in Digital Pathology accelerating– Europe leads United States
– Companies looking to go all digital
Partly cloudy outlook– Cloud providers partnering with med tech
Age of informatics - analytics platforms– Delivering insights from vast quantities of data
– Indica Labs, Inspirata, Proscia
Data is being shared and retained longer– Larger images and more complex image analysis
– Elixir, National Data Services
9/6/2016
Long Live Data™
6
HGST in Your Lab
©2015 HGST, Inc. All rights reserved. Confidential
Local workstations– Sequencers & Microscopes
– Flash, HDD & SSD
– Value: faster data/image capture and transfer
Processing server cluster– PCIe SSD for data acceleration
– Infiniflash for primary storage and buffering
– Value: analytics and acceleration, allows use of less specialized servers
Active Archive System & JBODS– Consolidate tier-2 and backup into single
shared archive
– Value: collaboration, cloud scalability, greater data durability
9/6/2016 7
Long Live Data™
©2015 HGST, Inc. All rights reserved.
Traditional storage solutions don’t fit data at scale
DIY storage is inefficient
Cloud storage costs are not predictable
Compliance and security concerns
Getting data to the cloud
More Data, More Challenges
9/6/2016 8
Long Live Data™
AWS Public CloudPrivate Cloud w/Bursting
Leverages the Power and Scale of Object Storage
9/6/2016 9
Long Live Data™
What Does 1 Petabyte Look Like? InfiniFlash System™
Ultra-dense High Capacity Flash storage– Up to 512TB in 3U, Scale-out software for PB scale capacity
Highly scalable performance–Industry leading IOPS/TB
Robust storage services –Automatic rebalancing
–Hot Software upgrade
–Snapshots, replication, thin provisioning
–Fully hot swappable, redundant
Integrated with a broad array of SDS platforms–Spectrum Scale, Red Hat, OpenStack + Ceph, Nexenta
9/6/2016
Long Live Data™
©2016 HGST, Inc. All rights reserved.
Built upon the software technology heritage of
Active Archive OS 4.1.1
• Rateless Erasure Code• 3 geo dispersion • Encryption – DARE• Usability improvements• BitDynamics®
• BitSpread™
Complete scale-up and scale-out object storage system
©2015 HGST, Inc. All rights reserved.
672TB-4.7PB raw capacity
Unbreakable Durability
Breakthrough TCO
Linear Scale Performance
Simplified Management
What Does 5 Petabytes Look Like? Active Archive System™
©2016 HGST, Inc. All rights reserved.
Thank You
Please Join Us Later for the Reception
9/6/2016 12
Long Live Data™
Value Comparison
Private CloudPublic Cloud
Pros– Single point of contact
One throat to choke
– Entirely Opex
– Operational expertise
– Broad developer support
– Dynamic compute scalability
– Ease of application deployment
Cons– Longitudinal data TCO -
– Significant egress costs
– Security and compliance concerns
Proof of secure erase and data locality
Pros– More cost effective than AWS
Longitudinal data TCO - $$$$$$$$$
Glacier cost structure w/std. cloud performance
– Local support and resources (SDSC)
– Primary customer; specialized experts directly involved
– Data locality and security
– AWS-like storage availability w/o replication
– Allows geo-dispersion among research partners (TACC, etc.)
– No ingress/ egress costs
Cons– Multiple partners
– Small developer community
– Ease of application deployment ???
9/6/2016
Long Live Data™
13
Challenges with Traditional Storage Architectures
©2015 HGST, Inc. All rights reserved. Confidential
Complex architecture with multiple tiers of expensive specialized storage
Limited scalability
Costly storage software licensing
Higher system management costs
Expense of onsite and offsite tape storage for disaster recovery
Overhead of tape media management
No protection from data degradation means no guarantee archives will be readable
9/6/2016
Long Live Data™
14
Customer Pain Points
©2015 HGST, Inc. All rights reserved. Confidential
Costly infrastructure does not scale (NAS and/or tape), hindering research
Infrastructure does not support global collaboration and shared file access
Complex infrastructure with data silos and diverse software apps
Unexpected storage requirements
Machines in constant use, more data is generated
What’s creating new opportunity?
Faster, cheaper instruments mean more scenarios investigated more data kept longer
New tools and analytics mean new insights from old data increasing data value
Technology advancements have downstream effects lens change = 2-5x more data per image
9/6/2016
Long Live Data™
Life Sciences Use Cases
Active Archive
Genomics
Bioimaging
Journal Archives
©2016 HGST, Inc. All rights reserved.
9/6/2016
Long Live Data™
16
Life Sciences Workflows
©2015 HGST, Inc. All rights reserved. Confidential
Genomics– Genome sequenced and raw data is generated,
highly compressible text data
– Multiple passes made to generate desired quality/accuracy
– Genome is reconstructed/aligned from data during processing phase
– Variants to reference genome identified and recorded
– Output: structured data file of variants
Bioimaging– Image data captured from slides (little
compression)
– Image processing and distortion correction
– Analyze image to measure size, composition, etc.
– Output: image composites and measurement data
– 3D and cryo-electron microscopes mean giga- and tera-pixel images
Similar but different