The State of Cloud Storage in 2013

THE STATE OF Cloud Storage IN 2013

On June 1st 2012 the number of
objects stored in Amazon S3 reached
1,000,000,000,000 one trillion

OR

142 objects
for every person on Planet Earth
3.3 objects for every star in
our Galaxy

If you could count one object per second it would _. take you 31,71 0 years to count them all

DATA SIZE

It is estimated there is over 1 Exabyte of data stored in the cloud”
This is over 67 million iPhones worth of data

At year-end 2016, more than 50 percent of Global 1000 companies will have stored customer-sensitive data in the public cloud
- Gartner, Top Predictions for IT Organizations and Users for 2012 and Beyond
OR 1,073,741,824 Gigabytes

WHAT DOES THIS MEAN FOR IT?
As the public cloud becomes more integrated with
enterprise infrastructure, independent testing and
monitoring of these systems becomes more critical to IT

?
CSPs

To this end, there are three major benchmarks for the Cloud Storage Providers (CSPs)
Performance
Availability
Errors

WHAT SIZE FILES ARE USERS STORING?
File size distribution of files stored by the typical enterprise user

1 KB 10KB
100 MB 1 GB 0.4% 0.1%
1MB 100KB
9.7% 16.8% 22.2% 24.6% 26.2%

Over 75% of files stored are less than 10MB in size


THE BENCHMARKS TESTS


1. PERFORMANCE How fast can data be read/written to the cloud?

THREADCOUNT 50, IN MB/S, LONGER BARS ARE BETTER

TO WRITE 10 KB
Microsoft ----- WINNER
Amazon
HP
Rackspace
Google

TO READ 10KB
Microsoft ------ WINNER
HP
Amazon
Google
Rackspace


TO WRITE 10MB
Rackspace ---------------------WINNER
Google
Amazon
Microsoft
HP

TO READ 10MB
Microsoft ---------------WINNER
Rackspace
Amazon
HP
Google


2. AVAILABILITY Is data stored in the cloud always available (read)?
Is the cloud available to accept data (write)?

AVERAGE AVAILABILITY RESPONSE TIME, IN SECONDS

TOP PERFORMERS
Microsoft
Amazon
Rackspace

2.0 Google
1.0 HP
Rackspace
Microsoft
Amazon

0 Nov1 5 10 15 20 25 30

Testing was conducted from Nov 1st 2012 to Nov 30th 2012



3. ERRORS As the number of objects stored in the cloud increases are there resulting errors?

PERCENTAGE OF WRITE ERRORS
Microsoft 0%
Google 0%
Amazon 0%
Rackspace 0.0000001%
HP 0.000017%


PERCENTAGE OF READ ERRORS
Microsoft 0%
Rackspace 0.0012%
Amazon 0.0018%
Google 0.0030%
HP 0.0099%

Microsoft was the only provider to achieve zero read and write errors

Even the CSP with the highest error rate performed well. At 0.000017% you would be 500X more likely to be struck by lightening than experience a write error

HOW DID THEY STACK UP?

1st MICROSOFT
2nd AMAZON

Read the full report at www.nasuni.com/cloudreport THE STATE OF Cloud Storage IN 2013 On June 1st 2012 1,000,000,000,000 the number of objects stored in Amazon S3 reached one trillion OR 142 objects for every person on Planet Earth 3.3 objects for every star in our Galaxy If you could count one object per second it would take you 31,710 years to count them all Source: http://aws.typepad.com/aws/2012/06/amazon-s3-the-first-trillion-objects.html -...DATA SIZE It is estimated there is This is over At year-end 2016, more than 50 percent of Global 1000 companies will have stored customer-sensitive data in the over 1 Exabyte of 67 million iPhones 16GB iPhone! data stored in the "cloud" worth of data public cloud OR 1,073,741,824 Gigabytes - Gartner, Top Predictions for IT Organi- zations and Users for 2012 and Beyond .WHAT DOES THIS MEAN FOR IT? WHAT SIZE FILES ARE USERS STORING? .... File size distribution of files stored As the public cloud becomes more integrated with enterprise infrastructure, independent testing and monitoring of these systems becomes more critical to IT by the typical enterprise user 1 KB 10 KB .**************..... 16.8% 100 MB 0.4% 24.6% 1 GB 0.1% 10 MB 22.2% 9.7% 26.2% ...... CSPS 1 MB 100 KB Performance To this end, there are Over 75% of three major benchmarks for the Cloud Storage Availability files stored are less Providers (CSPS) Errors than 10MB in size .THE BENCHMARKS TESTS 1 PERFORMANCE How fast can data be read/written to the cloud? THREADCOUNT 50, IN MB/S, LONGER BARS ARE BETTER TO WRITE 10KB 10MB MB/S MB/S 8 30 60 90 120 E Microsoft WINNER! Rackspace WINNER! I Amazon 3 Google hp HP Amazon A Rackspace Microsoft 2 Google hp HP TO READ 10KB 10MB MB/S MB/S 5 10 15 20 40 80 120 160 Microsoft Microsoft WINNER! WINNER! hр НP A Rackspace I Amazon I Amazon Google hp HP O Rackspace 3 Google 2 AVAILABILITY Is data stored in the cloud always available (read)? Is the cloud available to accept data (write)? AVERAGE AVAILABILITY RESPONSE TIME, IN SECONDS TOP PERFORMERS 2.0 Google #1 Microsoft НР Rackspace Microsoft # 2 Amazon Amazon .0 Nov.1 10 15 20 25 30 #3 Rackspace Testing was conducted from Nov 1st 2012 to Nov 30th 2012 3 ERRORS As the number of objects stored in the cloud increases are there resulting errors? PERCENTAGE OF WRITE ERRORS PERCENTAGE OF READ ERRORS A Microsoft 0% F Microsoft 0% Microsoft was the only provider to achieve zero 3 Google 0% Rackspace 0.0012% I Amazon 0% I Amazon 0.0018% read and write errors O Rackspace 0.000001% Google 0.0030% hp HP 0.000017% пр НP 0.0099% Even the CSP with the highest error rate performed wellI. At 0.000017% you would be 500X more likely to be struck by lightening than experience a write error HOW DID THEY | ST MICROSOFT STACK UP? 2ND AMAZON nasuni. Read the full report at www.nasuni.com/cloudreport SECONDS