Storage

Changing the rules of storage

Share this post:

Over the decades we have become masters at storage. With each technology advancement and approach across file, block and object storage–we have found new ways to sock away ever increasing volumes of data and content.

As the adoption of cloud has taken hold in our personal and professional realms, we are now able to sock away seemingly infinite volumes of data.

And cloud providers have come up with an ever increasing number of options and models  to handle practically every permutation of storage requirements. But that has come with associated complexity. Especially when you consider that in business you do not have a monoculture of storage needs. Your organization has diverse needs that dictate:

  • Whether your data and applications need to be in your data center or if they can take advantage of cloud.
  • Whether your data is hot — meaning that you expect to access it quite frequently from your cloud and mobile applications; cool — meaning that you are in effect backing up or archiving data that you do expect to access albeit infrequently; cold — meaning that you expect that you will not have to access the data, potentially ever.

All of that assumes you have close to absolutely certainty about your data and its usage patterns in the future.  That may be possible if we were talking about one application and associated data sets, maybe 5 or 10, but what happens when you are looking at petabytes of data that supports a myriad of applications?

Figure 1 Typical Cloud Storage Pricing Model

Figure 1 Typical Cloud Storage Pricing Model

The cost of being wrong has never been higher — even when you take advantage of cloud storage. Figure 1 is an atypical, (actually one of the less complicated) cloud storage pricing model. It requires that you know how much you are storing, and whether you expect that data to be hot, cool or cold in the future. Add to that the complexity of selecting from multiple regions for your storage, and the complexity increases significantly.

That creates complexity and risk. Two things no CIO, CFO, IT Architect or even developer enjoys hearing.

Figure 2 IBM Cloud Object Storage Flex Pricing (Draft pricing. Final pricing at availability)

Figure 2 IBM Cloud Object Storage Flex Pricing (Draft pricing. Final pricing at availability)

With IBM Cloud Object Storage Flex (FLEX), we took a different approach. We started by asking a simple question – why do we have storage “tiers” in the first place? Tiers are certainly great when you posess close to absolute certainty of your future usage patterns. But business today is uncertain. Opportunities are fleeting. FLEX is ideal for the situations where your data usage patterns are unpredictable or indeterminate. FLEX addresses the uncertainties you face in your business without the hindrances of traditional tiered pricing models. And this innovative technology, which is supported by over 400 patents, is being offered at a significantly lower price point than our top competitors.

Figure 3 Comparing IBM Cloud Object Storage

Figure 3 Comparing IBM Cloud Object Storage

Figure 3[i]  shows results of in-house analysis. On months where data runs hot and you need to readily access it — FLEX is close to 24% at ½ petabyte and 22% at 5 petabytes less than AWS storage options.

On months where data runs cold and you don’t access it frequently — FLEX promises to be a whopping 53% less!

And for Microsoft Azure? Around 25% at ½ petabyte and close to 23% less at 5 petabytes on hot usage months and 75% less on cold usage months. Those type of numbers should have you interested.

But there is a bigger vision at play.

With the simplified and streamline pricing that FLEX  will deliver- we stand to make it significantly  easier for clients to shift from merely storing data to more readily using  data at will to  derive insights with our advanced cognitive and analytics portfolio. Our vision is one where you can quickly employ cognitive and analytics against the volumes of data you are storing. You can access your data in 100s of milliseconds speed and it is protected to 11 9s — and it is ready to go. The use cases for this compelling! Here are just a few to consider:

  • Genomics and research — You may be storing genomic data to cure a rare disease. As you collect mountains of genome data, you archive it in FLEX. But when you need to access it to run Watson visual recognition against petabytes of content, you can easily do that with FLEX and the IBM Cloud and Watson platform and not pay exorbitant penalties.
  • Energy exploration and analytics — You may be storing seismic data from exploration rigs around the world. You are archiving that data but you want to bring it on line fast to merge it with a weather data set using SPARK. With FLEX, the data is there when you need yet still protected and secure.
  • Manufacturing — You’ve just brought a new modern factory online and your automation is generating tons of IOT data that you want to capture and utilize to improve workflow and yield. This data is needed for long term studies but some if it may be very active during certain cycles of processing. FLEX is perfect for this highly variable workload that will help you produce your products more efficiently.

One thing is certain. The future is uncertain. We really don’t know how we may use the volumes and volumes of data we have amassed. But what we do know is:

  • We want the data secured to 11 9s
  • We want protection against regional outages
  • We want the data, when we want it and do not accept multi-hour or multi-day delays in getting our data online from cold storage
  • We want to control our costs
  • We see data and content as essential ingredients that with cognitive and analytics can help us gain a competitive advantage

FLEX is expected to be released in early Q2’ 2017 in the US. For more information, please visit ibm.com/cloud or bluemix.com.

In addition to FLEX, we announced a number of other elements that extend and add value to our IBM Cloud Object Storage family.

  • IBM Cloud Object Storage Cold Vault and Vault are now available via digital delivery from Bluemix. These two offerings are ideal for long term archival and infrequent access respectively.
  • IBM Cloud Object Storage via IBM Bluemix Garage. The IBM Bluemix Garages around the world provide a unique space for clients to explore how object storage can be combined with the broader IBM Watson and Cloud platform to create rich, engaging and transformative applications.

NetApp Integration.  AltaVault Backup accelerates enterprise hybrid cloud adoption by extending workloads seamlessly into the IBM Cloud; enables clients to leverage the NetApp Data Fabric to unlock the value of data through IBM Watson cognitive capabilities.

[i] Figure 3 comparison

For AWS – Comparison: IBM Cloud Object Storage Flex Cross-Region vs. S3 “Hot”: Standard bucket in AWS US East (“primary”) with Cross Region Replication to S3 Infrequent Access bucket in US West Oregon (“replica”) and vs. S3 “Cool”: S3 Infrequent Access bucket in AWS US East with Cross Region Replication to S3 Infrequent Access bucket in US West Oregon. All S3 access assumed through “primary” bucket.  Pricing based on Amazon US list prices as of 3/6/2017 and announced IBM Cloud Object Storage “Flex” pricing with planned availability for April 2017.  Price includes storage capacity, API operations, and cross-region data replication charges (S3 only).  Price comparisons exclude outbound internet data transfer charges. Usage based pricing comparison based on specific workload assumptions – actual customer charges will vary depending on workload, storage capacity, object sizes, data access patterns, storage tier, redundancy options, and other factors.  Pricing for this comparison based on the following workload assumptions:

  • Mixed footprint of 5% “small” and 95% “large” objects (by capacity), Average object sizes:  Small = 3MB, Large = 275MB
  • Monthly access patterns: “Hot”: 80% read , 50% written, 5% listed.  “Cool”:  10% read , 50% written, 5% listed.  All objects assumed retained at least 30 days.

IBM Cloud Object Storage

Amazon AWS S3:

Cross-Region Replication

Amazon Simple Storage Service (S3) FAQs

Amazon S3 Pricing

For Azure  – Comparison: IBM Cloud Object Storage Flex Cross-Region vs. Azure Hot GRS in East US 2 (Primary) and vs. Azure Cool GRS in East US 2 (Primary).  Pricing based on Azure US list prices as of 3/6/2017 and announced IBM Cloud Object Storage “Flex” pricing with planned availability for April 2017.  Price includes storage capacity, API operations, and data replication charges (Azure only).  Price comparisons exclude outbound internet data transfer charges. Usage based pricing comparison based on specific workload assumptions – actual customer charges will vary depending on workload, storage capacity, object sizes, data access patterns, storage tier, redundancy options, and other factors.  Pricing for this comparison based on the following workload assumptions:

  • Mixed footprint of 5% “small” and 95% “large” objects (by capacity), Average object sizes:  Small = 3MB, Large = 275MB
  • Monthly access patterns: “Hot”: 80% read , 50% written, 5% listed.  “Cool”:  10% read , 50% written, 5% listed.

IBM Cloud Object Storage:

Azure GRS:

Azure Storage Pricing

Azure Blob Storage: Hot and cool storage tiers

 

More Trending Stories

99.95% availability. Balancing release velocity and reliability

Availability and reliability are rarely at the front of developers minds when delivering new applications on Bluemix. The ease and speed of creating and deploying new features is very seductive.

Continue reading

IBM Cloud Storage Hourly Billing: Disaster Recovery and Dev Ops Options

IBM Cloud Infrastructure team is delivering new features and functions to you at an accelerated pace. Today, IBM announces hourly billing for its Cloud Block and File Storage services. This added feature pairs well with IBM’s hourly compute instances. Imagine – you can create a compute instance with storage and pay for both by the hour! Now you have the choice – monthly or hourly. The IBM Cloud aims to make infrastructure flexible to meet customers’ needs.

Continue reading

Scale databases on-demand with Db2 on Cloud

IBM Db2 on Cloud (formerly dashDB for Transactions) has released a scalable plan that allows customers to scale RAM/CPU and disk independently.

Continue reading