Infrastructure Wars: The Battle Brewing in the Storage Industry

How Google, Microsoft and Oracle are Driving Competition in the Storage Industry 

Application Stack vs Array

 

What you Need to Know 

There is a competitive battle brewing in the on-premise storage business and it’s not between EMC/NetApp or EMC/IBM. It’s stemming from a move by independent software vendors specifically Microsoft and Oracle, to bundle more storage function into their application stacks, push storage function closer to the host and commoditize the storage hardware layer. The move to integrate storage function into the application stack is real and in some cases can add substantial value to organizations. But there is a price to pay and IT executives need to understand the strategies and implications for long term success. Underpinning these trends is Google’s decade long march toward simplification and cloud services; which is not only driving software vendors like Microsoft crazy; it’s also causing them to drive down perceived costs wherever possible and grab as much value in their stacks as they can. 

Here’s the bottom line. IT execs have three choices: 

  1. Do Nothing: Don’t implement any coherent infrastructure strategy and allow application heads to do whatever they want (Bespoke Model). If you’re an organization that has excess cash to burn and is not interested in improving efficiencies or aligning IT infrastructure with business requirements this is a perfectly viable strategy for you.
  2. Standardize: Implement chargebacks and offer application owners a limited set of standardized, menu-based infrastructure services (Shared Storage Services Model) that are managed, evolved and best-of-breed. This strategy will piss lots of people off within the organization but you will end up with lower costs and a far better return on infrastructure assets.
  3. Compromise: Use a hybrid approach that defaults to a Storage Services Model but selectively integrates function into the application stack when warranted based on a clear business case (Hybrid Model).

Based on recent research conducted by the Wikibon community, we believe the Hybrid Model will yield the best business returns, integrating storage function into the application stack where it drives revenue or simplifies recovery; while at the same time leveraging the superior asset utilization dynamics of a Shared Storage Services Model. 

Background 

The history of the information technology industry can be described using a metaphor of waves that coincide with major technology cycles. The figure below depicts the four great eras of computing: Systems-centric (dominated by the mainframe), PC-centric, network-centric and now service-centric, specifically Web services. 

Riding the Google Wave

 

Last summer, in conversations with a number of Wikibon members, we identified a key trend that caught our attention. In response to the economic crisis, CIO’s, Application Owners and Infrastructure Executives were collaborating on ways to simplify infrastructure and make their operations more cost effective and elastic. Our belief is that the consumerization of IT generally and specifically Google have changed the way in which senior IT and business managements approach technology decisions. 

We kicked off a research initiative to study this trend further and its specific impact on storage. We conducted 18 interviews with IT practitioners in finance, insurance, banking, retail, manufacturing, health care and technology. Our research mainly focused on Oracle and Microsoft shops with additional research into Oracle 11g and Microsoft Exchange 2010. The main goals of the project was to: 

  1. Understand how CIO’s, Application Owners and Infrastructure Heads make decisions about infrastructure;
  2. Understand and quantify the value of two different approaches to infrastructure: 1) Placing storage function in the application stack or 2) Using a Shared Storage Services Model.

Key Findings: Implications for Storage Companies 

Several research notes emerged from these projects (outlined below). The key findings contain both threats and opportunities for  storage vendors, specifically: 

  • The cold hard truth is that certain storage function is steadily migrating to the application stack for good reason. An example is Oracle’s Automatic Storage Management. It’s bundled, integrated and users like it. Oracle and Microsoft are intent on capturing more value and storage function is a clear target.
  • Microsoft (Exchange 2010), and often Oracle, are recommending its customers don’t use SAN and many Application Owners are paying attention.
  • This strategy is in many cases being driven by Google’s steady move toward cloud services; which threaten Microsoft and are forcing it to create the perception of lower costs by, for example,  recommending DAS for Exchange 2010.
  • The good news for  storage suppliers is:
    • While Oracle is formidable and is very good at database recovery, Microsoft’s Exchange storage stack offering is deficient and exposes customers to certain risks. In particular, Microsoft’s IP-based replication approach is not advisable for many customers due to its inherent inability to guarantee consistency.
    • Using DAS for Oracle applications is almost always more expensive due to the high cost of Oracle licenses and the lack of asset leverage in pure DAS environments.
    • In Microsoft Exchange environments with more than 1,000 mailboxes, SAN is consistently cheaper than DAS.
    • Lower cost from SAN in Oracle 11g and Microsoft Exchange come from better asset leverage achieved from SAN. Simply put, SAN is a re-usable asset across multiple applications and DAS isn’t. As such, DAS, combined with Oracle or Microsoft storage function, is more often than not, a significantly more expensive proposition for mid-to-large enterprise customers.
    • Virtualization is a key driver in the organizations we researched and a shared storage infrastructure is more advantageous in these environments.

The bottom line is that Microsoft and Oracle are competing aggressively to capture as much storage value as possible. This trend will likely continue for the foreseeable future and Oracle in particular has a very strong story in its database markets. However a shared storage services model will be more cost effective for the vast majority of enterprise customers. 

Here’s a chart that underscores this finding by comparing a Shared Storage Services Model to a “One Off” Bespoke approach. More detailed explanations can be found in the research notes below: 

Comparing Shared Storage Services Costs to One Off (Bespoke)

 

What this Means to IT Executives 

This research has several implications for different IT roles. The nature of chargebacks will largely determine the degree to which an organization is willing to mandate a shared storage services model. Organizations with chargebacks consistently indicated that one-off or bespoke implementations, including software stack approaches, were only tolerated when the business value was clear. Otherwise, these organizations would default to a Shared Storge Services Model. Consistently these organizations reported more efficient use of IT infrastructure. In other words, chargebacks are a CFO’s friend. 

Organizations without chargebacks consistently indicated to Wikibon that application owners dictated the infrastructure strategy and pretty much get what they want, when they want it. 

Our advice to IT executives is as follows: 

CIO’s 

Your application groups are often pursuing strategies that incur unneccessary expense. There is clear proof in the research we conducted that using a Shared Storage Services Model to manage infrastructure investments will allow you to achieve substantially better returns on your assets by: 1) driving standardization and 2) leveraging infrastructure across your application portfolio. 

As application heads pursue more expensive infrastructure strategies ask a simple question: Is this extra expense justified? Does it drive enough business value to offset the higher costs of the one-off (bespoke) strategy? If so, great, go for it. But our research indicates that in the vast majority of cases, you’ll get more business value out of a shared infrastructure approach. 

CIO’s should focus on delivering business value and avoiding risky decisions that hurt the health of the organization. So be careful to have vendors demonstrate how to get from where you are today in a Bespoke situation to a Shared Storage Services Model, without taking on all the risk. 

Application Heads 

In most cases, you the application head thinks about cost of infrastructure last. But the smart thing is to default to a Shared Storage Services Model  unless bundling function into the application delivers clear business value. Oftentimes this is the case, for example with Oracle ASM. But in many cases, application heads are implementing platinum infrastructure service where ‘good enough’ would suffice. By leveraging a standardized infrastructure you are applying similar concepts to SOA which will speed time to deployment, lower costs and deliver more business value. The bottom line is a horizontal Shared Storage Services Model that services more applications in the portfolio is the best strategy for most of your applications. 

Obviously you need storage vendors to prove this to you. Service levels are king and you shouldn’t do anything to compromise them. Our contention is that in many cases, a Shared Storage Services Model will yield better service levels and a better end user experience in terms of performance and recovery. This is not always the case and that’s where an application stack approach makes sense. 

But don’t let politics (e.g. “I don’t want to get storage services from a separate group; I want to control my own infrastructure”) get in the way of sound business decisions. If it makes business sense to integrate storage function into the application stack then do it. But don’t just blindly trust the ISV or the storage vendor – do your own homework and make the right business decision based on your overall objectives. 

Infrastructure Heads 

Infrastructure Heads at mid-to-large organizations generally buy in to the shared storage philosophy. The key for you  is to arm yourselves with data and proof points that you can communicate to the CIO, CFO, Infrastructure Heads and other business line managers. 

Key Takeaways 

In summary, the Google effect is impacting the behavior of both buyers and sellers. CEO’s are observing the simplicity of cloud computing and driving top-down edicts to reduce IT complexity; which is rippling through to infrastructure heads. For customers not outsourcing to the cloud the two prominent models are a ‘bespoke’ strategy placing function into application stacks versus keeping function in the array. The former is more expensive and the latter causes political tension related to forcing a fixed services menu on application heads and business owners. CIO’s need to cut through the politics and make infrastructure decisions based on business value. If the business derives value by placing function in the application stack it should be done so aggressively. Otherwise a shared storage services model will be more efficient. 

Key Takeaways

 

Published Research 

To assist IT professionals in making better decisions, we’ve published several research notes and blogs. The following related research items can be used to support your goals: 

  1. Why Microsoft’s Head is up its DAS – a summary of how Google is forcing Microsoft’s hand and pressuring the company to do anything to lower the perceived cost of Exchange, including recommend DAS exclusively.
  2. How Google Apps Threatens Microsoft – the detailed research behind the blog post (#1) with a cost analysis of Exchange on-premise, Exchange On-line and Google Apps.
  3. Should Exchange 2010 use DAS or SAN? An economic analysis of DAS vs SAN in Exchange 2010 environments; including a view of where costs cross over.
  4. Should Storage Services Reside in Arrays or Application Stacks? An objective view that lays out the two broad strategies for placing storage function with a specific drilldown into Exchange 2010 and Oracle 11g.
  5. A Storage Services Approach will Streamline Infrastructure Delivery: A definition of a shared storage services model with a financial assessment comparing the total cost of ownership of shared storage vs. a One-off (bespoke) orientation.
  6. Comparing Return on Assets (ROA) to  Return on Investment (ROI) – The Wikibon ROA model underpins the financial analysis used in this research. The model specifically evaluates the value of installed assets over a period of time and is particularly useful in assessing asset utilization across an application portfolio.

Additional Research 

  1. A Discussion with Practioners: Applications or Infrastructure-centric Approaches to Functional Delivery
  2. Should Function Reside in Infrastructure or Application Stacks?

As always, we appreciate your participation as a Wikibon member to help your Peers make better IT decisions. We hope you find this information useful and we look forward to your feedback and assistance with future initiatives.

Share

, , , ,

  • Pingback: uberVU - social comments()

  • Pingback: Most Tweeted Articles by Storage Experts Experts: MrTweet()

  • Alma Cenada

    If I am choosing storage for inside my company (not cloud), then I think about history of networked storage is a way to share storage so I can buy it and manage it more efficently. This makes easy sense to me. So if storage in the stack is now better, that means stack owner was making storage better more quickly than networked storage vendor — you give Oracle ASM example. That is too bad I think. But I see also some network storage vendors are making new advances, so I am not sure. But for my friend's business which is small I told him DAS.

  • tony08

    The last two locations that I worked at the san team itself was way more costly then just having das storage laying around being wasted. So it did'nt matter that san made the storage more available for other applications it was the FTE cost of the storage team that killed things. DAS is quite a bit simpler to manage and you don't need dedicated staff for storage.

  • dvellante

    Interesting comments Alma and Tony. My experience is at scale SAN is cheaper than DAS. But it is more complex and requires more experienced and expensive FTEs to manage. So like many things the answer it it depends.

  • Pingback: Practitioners Discuss Infrastructure Strategies « Wikibon Blog()

  • Pingback: Stack Wars: The Links – Gestalt IT()

  • Pingback: Oracle’s New ZS3 Underscores Integration End Game « Wikibon Blog()