© 2009 IBM Corporation –
IBM XIV Storage System
Patrick Lee
Enterprise Storage Reinvented
Cert if ied Consult ing Systems Specialist
October, 2009
2 © 2009 IBM Corporation
XIV Company Profile
Execut ive Chairman - Moshe Yanai–Invented and led Symmetrix systems development
Founded in 2002 by a group of 5 graduates of Talpiot–Talpiot is an elite Israeli academic program for the sciences,
physics, and mathemat ics–All f rom the 14th year of the proj ect (XIV=14 in Roman numerals)
Product in development for 6 years–More than 50 patents f iled (maj or architectural patent has
already been approved)1000+ systems shipped to date worldwide–90+ % of evaluat ion systems have gone into product ion–Over 40% of Fortune 500 have deployed XIV–A/ NZ growth from 0 to 12 systems in 2009 so far
3 © 2009 IBM Corporation
XIV now –an IBM Company
XIV was acquired by IBM on December 31, 2007XIV is now operat ing as a subsidiary of IBM report ing to IBM’ s storage organizat ionMoshe Yanai was appointed an IBM FellowFor our customers, this means:–Next -generat ion storage product–IBM support and services
© 2009 IBM Corporation
With this legacy architecture, scalability is achieved by using more powerful (and more expensive) components
Current Enterprise Storage Solut ions
Cache
Controllers
Interface Interface InterfaceBuilding blocks:• Disks• Cache• Controllers• Interfaces• Interconnects
JBOD JBOD
Scale Up
PERFORMANCERELIABILITYSCALABILITY
= $$$
©20
09 IB
M C
orpo
ratio
n
Trad
itio
nal S
AN M
anag
emen
t
••All
disk
s acc
essib
le b
y al
l con
trol
lers
All d
isks a
cces
sible
by
all c
ontr
olle
rs
••Cac
he m
irror
ed
Cach
e m
irror
ed
••Tie
red
stor
age
by ty
pe
Tier
ed st
orag
e by
type
••Hot
spar
es n
eede
d by
type
Hot s
pare
s nee
ded
by ty
pe
••Seg
men
t by
RAID
type
& se
ts
Segm
ent b
y RA
ID ty
pe &
sets
••Spa
re sp
ace
Spar
e sp
ace
disc
ontig
uous
disc
ontig
uous
••Disk
pla
cem
ent m
atte
rs
Disk
pla
cem
ent m
atte
rs
••Pre Pre--
plan
for g
row
th
plan
for g
row
th
••Tun
e Tu
ne L
UNs
LUNs
for s
egm
ent s
ize
and
owne
rshi
p
for s
egm
ent s
ize
and
owne
rshi
p
••Mor
e fin
e tu
ning
More
fine
tuni
ng……
..
••How
do
I mai
ntai
n th
is as
dat
a gr
ows u
neve
nly?
How
do
I mai
ntai
n th
is as
dat
a gr
ows u
neve
nly?
© 2009 IBM Corporation
IBM XIV Storage Architecture
Interface Interface
Design principles:• Massive parallelism• Granular distribution• Off-the-shelf components• Coupled disk, RAM and CPU• User simplicity
Data Module
Interface Interface Interface
Data Module Data Module Data Module Data Module Data Module Data Module
Switching Switching
Scale Out
© 2009 IBM Corporation
Data Module
UPSUPS UPS
Data Module
Data Module
Data Module
HostHost
HostHost
FC
Ethernet
iSCSI
HostHost
XIV System Components
© 2009 IBM Corporation
XIV System Descript ion� 15 modules, each with:
– 12 x 1TB disk drives– 8GB memory– 1 Quad core CPU (2 on Interface Modules)
� 180 disk drives in each rack– (180TB raw, 79TB user data)
• Global spare space –full module plus 3 disks• 79TB = (180 –12 –3 )/ 2 –3.5 (internal use)
� 120GB of memory
� 24 FC ports
� 6 iSCSI ports
� 84 Intel Xeon cores� IBM Service and Support
– 1 year or 3 year, 4 hour response, 24x7 Same Day On-Site Repair
– IBM HW and SW Installat ion
© 2009 IBM Corporation
IBM XIV Storage Dist ribut ion Algorithm
� Each volume is spread across all drives
� Data is “ cut ” into 1MB “ part it ions” and stored on the disks
� XIV algorithm automatically dist ributes part it ions across alldisks in the system pseudo-randomly
Data Module
Interface
Data Module
Interface
Data Module
Interface
Switching
XIV disks behave like connected vessels, as the distribution algorithm aims for constant disk equilibrium.
Thus, IBM XIV’s Storageoverall disk usage could
Approach 100% utilization when loaded
© 2009 IBM Corporation
XIV Dist ribut ion Algorithm on System Changes
� Data dist ribut ion only changes when the system changes– Equilibrium is kept when new hardware is added
– Equilibrium is kept when old hardware is removed
– Equilibrium is kept after a hardware failure
Node 2
Node 3
Node 1
Data Module 2Data Module 1
Data Module 3
© 2009 IBM Corporation
Node 4
XIV Dist ribut ion Algorithm on System Changes
� Data dist ribut ion only changes when the system changes– Equilibrium is kept when new hardware is added
– Equilibrium is kept when old hardware is removed
– Equilibrium is kept after a hardware failure
Data Module 2
Data Module 3
Data Module 1
[ hardware upgrade ]
Data Module 4
© 2009 IBM Corporation
XIV Dist ribut ion Algorithm on System Changes
� Data dist ribut ion only changes when the system changes– Equilibrium is kept when new hardware is added
– Equilibrium is kept when old hardware is removed
– Equilibrium is kept after a hardware failure
Data Module 2
Data Module 3 Data Module 4
Data Module 1
[ hardware failure ]
The fact that distribution is full and automatic ensures that all spindles join the effort of data re-distribution after configuration change.
Tremendous performance gains are seen in recovery/optimization times thanks to this fact.
© 2009 IBM Corporation
IBM XIV Storage: Concept of “ Spare”
� Tradit ional approach– Dedicated disks used for spares
– In many systems spares are dedicated for a RAID group
� IBM XIV Storage approach– Recovery t ime: 30 minutes for 1 TB disk (full)
– No dedicated spare disk, only global capacity
– All disk are equally used
– Minimize the risk of technician mistakes
– Higher availabilit y with no performance impact
� 180TB raw is 79 TB net– Spare space for 3 disks and a full module
– 80 = (180 –12 –3 )/ 2 –3.5 (internal use)
© 2009 IBM Corporation
IBM XIV Storage: Thin Provisioning
� Defining logical volumes bigger than physical capacity
� Installing physical capacity only if and when needed
� No space consumed when data is 0
� Pools are used to manage quota
� Results:– Reduced overall direct storage cost
– Storage expenses spread over t ime, exploit ing price reduct ions
– Easier management
– Save 20-50% of storage capacity
© 2009 IBM Corporation
Data ModuleData Module
Data Module
IBM XIV Snapshots - Virtually without Limits� Snapshot creat ion/ delet ion is instantaneous
� High Performance WITH snapshots
� Unlimited number of snapshots
� Different ial snapshots save 15-30% of storage capacityVolume
Vol
Snap
Snap
Distributed snapshots on each module. Extremely fast memory operations
Accessing snapshots is as fast as accessing production volumes
As Host writes data, it is placed randomly across system in 1MB chunksEach module has pointers in memory to
the disks that hold the data locally
On a snapshot, each module simply points to original chunks. Memory only
operation
Restore Volume from snapshot copy
High performance snapshots provide:
• Easier Physical Backup to Tape• Instant recovery from Logical Backup• Easy creation of Test Environment• Boot-from-SAN with easy rollback• Easy Data-Mining on Production data
© 2009 IBM Corporation
IBM XIV Storage: Monitoring
© 2009 IBM Corporation
XIV Part ial Rack Configurat ions
• All options available to order• Field upgrade possible
© 2009 IBM Corporation
XIV Part ial Rack Configurat ions
© 2009 IBM Corporation
XIV Footprint
© 2009 IBM Corporation
DisclaimerCopyright © 2008 by Internat ional Business Machines Corporat ion.
This publicat ion is provided “ AS IS.” IBM product informat ion is subj ect to change without not ice.
No part of this document may be reproduced or t ransmit ted in any form without writ ten permission from IBM Corporat ion.
Product data has been reviewed for accuracy as of the date of init ial publicat ion. Product data is subj ect to change without not ice. This informat ion could include technical inaccuracies or typographical errors. IBM may make improvements and/ or changes in the product (s) and/ or programs(s) at any t ime without not ice. The informat ion provided in this document is dist ributed “ AS IS” without any warranty, either express or implied. IBM EXPRESSLYDISCLAIMS any warrant ies of merchantability, f it ness for a part icular purpose OR INFRINGEMENT. IBM shall have no responsibility to update this informat ion. IBM products are warranted according to the terms and condit ions of the agreements (e.g., IBM Customer Agreement , Statement of Limited Warranty, Internat ional Program License Agreement , etc.) under which they are provided.
Informat ion concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products in connect ion with this publicat ion and cannot confirm the accuracy of performance, compat ibility or any other claims related to non-IBM products.
IBM makes no representat ions or warrant ies, expressed or implied, regarding non-IBM products and services, including those designated as Server Proven.
IBM may not offer the products, services or features discussed in this document in other count ries, and the informat ion may be subj ect to change without not ice. Consult your local IBM business contact for informat ion on the product or services available in your area.
All statements regarding IBM future direct ion and intent are subj ect to change or withdrawal without not ice, and represent goals and obj ect ives only. Contact your local IBM off ice or IBM authorized reseller for the full text of the specif ic Statement of Direct ion.