Cloud as Competition for the Data Center – Part 3
Part 3 of 3: What Data Center Leaders Should Do About The Cloud Threat
Part 2 discussed why competition continues to appear for the data center. With this understanding of the prior threats we now approach what to do about this latest challenge.
How does cloud computing compare to these prior threats?
Cloud computing combines the strengths of the other challengers, and loses most of the inhibitors. It has a low cost of entry, simplicity, rapid responsiveness, and can be adopted selectively. At first glance, business executives will see this as a very appealing alternative to data center capital expenditure. It is a challenge that cannot be ignored.
How should the enterprise data center executive respond to the latest challenge? There are three steps you should take:
- Know your competition
- Become an IT service broker for your company
- Develop your strategy for hybrid capability
1. Know your competition:
There is a quick way for a data center executive to truly understand the competition provided by cloud computing. Try it. Create an account with Amazon Web Services (and suitable credit card) to get started. Then try getting a server up and accessible. These are my results:
- 02.01p – Decide to get a server
- 02.02p – Order a server (choose the o/s image type, location and size; create a key pair if this is your first time; create security group for open ports; and launch)
- 02.08p – Connect to server
The responsiveness of the cloud was demonstrable in less than ten minutes. The entry cost for experimenting with infrastructure is negligible. This test was rounded up to an hour of usage and cost $0.03. Now compare this to the service level delivered by your data center. Can you deliver service immediately? Can your business units pay only for what they use? Have you simplified compute infrastructure down to multiple-choice selections?
Yes, this is an unfair comparison. There are many other aspects to compare besides cost and responsiveness including availability, service levels, recoverability, and compliance (feel free to add more in the comments). The servers provided by data centers come customized to company policy with management and protection software pre-installed, not a bare bones image. Yet the comparison is compelling.
Cloud computing sets a new expectation for service from the data center.
Developing an equivalent internal service for your company will take time, but there is an interim step you can take.
2. Become an IT service broker for your company:
Data center executives should step into the cloud and quickly. If not already, your business will soon be using cloud servers and storage. You face a choice. You can be out of that loop, or you can step in as the expert broker in cloud negotiations. Negotiate a deal with Amazon, Rackspace, Verizon, IBM or myriad others that is cheaper than the rack rate your business can find on the internet. Market your alternate service internally … “have the hassle of dealing with Amazon directly or get the same service from me and save x%”
Of course, you need to deliver the same service level.
- Work with finance to offer this service as pay as you go.
- Make it a self service on demand offering – if it takes three weeks to get approval from you before your customers see the server, then the discount you add will not help.
You need to plan for how you can bring these services back into the data center later on. Your customers may realize the grass is not always greener on the other side.
Stepping in as a broker allows you to offer the service quickly, monitor the computing that is going on, and add value by identifying savings. It gives the data center the ability to compete immediately, and creates a window to create the plan to deliver such services internally.
3. Develop your strategy for hybrid capability:
Designs for internal cloud computing should be architected with eventual hybrid capabilities in mind. A simple internal cloud computing ability suffices to respond to the competitive challenge. But why stop there? Extending the goal to developing hybrid ability will allow the data center to offer a previously unavailable service, and strengthen the architecture.
Extending internal data center capacity is an expensive and time-consuming task that can obstruct a successful business initiative. The new business process captures customers’ attention and gains popularity causing performance to suffer under the load. Support falters while the data center procures and implements the needed capacity into the data center and the magic business moment is lost. Hybrid computing is an aspect of virtualization that promises to give us significant value. Sometimes called cloud bursting, it is the ability to add external computing capacity onto our existing internal ability transparently when needed.
Of course, this is more complex than it sounds. Which is why the hybrid approach acts as a good design test. The external and internal environments must be copacetic for this to work. Actually configuring this to work is a complex topic too deep for this blog. Nevertheless, this ability should be a basis for any internal cloud strategy.
The hybrid market is still unsettled having no standards and no clear market leaders. Thus the hybrid design will be limited. Even if you will never use external compute for security, compliance or other reasons, designing for it is a great criterion for ensuring you are exploiting the agility of the cloud. If you can go to an external cloud seamlessly, then it will be easy to elastically scale within your data center to cope with growth and equipment refreshes.
Will this response mean no future challenges to the data center?
No. As for the other challenges, cloud computing will mature and have to meet greater business expectations for availability, robustness, etc. Agility will eventually suffer as it mires in a complexity of options and choices. Hardware will become a constraining factor as we virtualize and consolidate open systems (cloud computing on an enterprise stack is a return to the mainframe model in many ways).
As enterprise data centers address each threat, we raise the level of service, find new efficiencies and add value to our relationship with the business units. Each threat is really a challenge that makes us stronger. Addressing cloud computing will take several years. We can already look forward to the next challenge!
Possibly Related Posts:
- Unified Communications Solutions
- Heat Index Reveals Hot Infosec Technologies
- The Ascent of 10GigE
- Storage Vendors See Mixed Q4
- High-risk Staff? Executives and IT Are Equally Risky
Tags: Cloud as Competition for the Data Center – Part 2
Part 2 of 3: Competition Exists When There Is a Better Choice
Part 1 revealed that over decades, the data center faced several competitive threats that shared certain traits. In order to understand what is happening today with cloud computing, we need to first ask a question.
…
Possibly Related Posts:
- Unified Communications Solutions
- Heat Index Reveals Hot Infosec Technologies
- The Ascent of 10GigE
- Storage Vendors See Mixed Q4
- High-risk Staff? Executives and IT Are Equally Risky
Cloud as Competition for the Data Center – Part 1
Part 1 of 3: Competition for the Data Center Is Not New
Staying competitive is the critical survival skill in market-based economies. Are data centers competitive? Data centers provide detailed and comprehensive computing services to the companies they support. These blog entries discuss why businesses consider alternatives to the data center, and what success cloud computing is likely to have as a competitor.
Cloud computing is not unique in competing with the data center. Since the first office was customized to support computers in the 1950s, the data center has faced challenges to its existence. Examining the common features of these alternatives has lessons to help us plan for and exploit IT as a service (XaaS) or cloud computing.
Competition for the data center arrives almost every decade. In the ’50s, mainframe computers starred in films as a godlike presence threatening Katherine Hepburn’s researcher job – they were untouchable (Watch “Desk Set” for amusement). Midrange or mini systems came along in the ’60s, promising that mainframes were going to disappear altogether (they did not).
Why were alternatives to mainframes appealing? Entry cost and agility. Mainframes required a significantly large investment that took time to install and configure (air conditioning, water cooling, special power and floors). Mainframe applications took even longer time to develop and deploy. Literally years could pass between your brilliant new business inspiration and having it become a service offered to clients. The frustrated business unit executive saw the data center as unresponsive and liked the idea of having a departmental computer instead. No longer would they have to compete with projects from other business units; they had control! And responsibility. Now they had to run a little data center, and soon they realized it was expensive and complex. Many of the departmental computers were eventually shipped back into the data center.
Another technology threat came from personal computers, and the centralized data center was announced dead (it is not). Again, the idea of personal control, easy access and short time to value appealed to departmental managers. But as PCs multiplied, managing all these devices and software running on them became an expensive overhead not related to their business. Commonly, the responsibility for personal computers moved back to the data center.
Later, outsourcing came along as a business threat to the data center. By this time, business executives learned they did not want to run data centers, they merely wanted to control the services. Outsourcing saw limited acceptance, as it required an all-or-nothing commitment of handing the data center over to someone outside the company. Competition from outsourcing led data centers to agree to service levels, helping both the business and the data center understand what the actual value of computing was to the company. (Note: Offshoring was not really a threat to the data center as it was simply about replacing workers with cheaper staff and not really a change in architecture.)
Why is there always a door open to a different architecture?
Check out Part 2 of Cloud as Competition for the Data Center for the answer.
About the Author
Marco Coulter, Managing Director of The Info Pro Cloud Computing Practice, utilizes nearly 30 years of experience in senior management and strategist positions on both the corporate and vendor sides of the IT industry. Marco’s expertise include storage, internal and external cloud, data center operations, software/hardware purchasing, software development, marketing, and product programs.
Prior to the TheInfoPro, Marco was an advisor to Enterprise Information Systems where he led the re-architecture of storage capacity and fabrics designing for a cloud-competitive future. Mr Coulter brings experience as a product line Vice President growing sales of the incubation project by 20% within one year while managing remote teams in India, Israel and China. Marco has held senior management positions in Storage Strategy and Product Management. Considered an industry expert in storage management technologies and international leadership, Marco is a sought after speaker at executive events worldwide.
Possibly Related Posts:
- Unified Communications Solutions
- Heat Index Reveals Hot Infosec Technologies
- The Ascent of 10GigE
- Storage Vendors See Mixed Q4
- High-risk Staff? Executives and IT Are Equally Risky
TheInfoPro’s 2010 Information Security Study Reveals Budget Changes, Cloud Concerns, Potential M&A Targets
- Larger vendors are leading in choice for infrastructure upgrades, points to potential M&A targets
- Forty percent (40%) of organizations are increasing security budgets in 2010
- Sixty percent (60%) of organizations already utilizing cloud-based infrastructure services or intending to do so in the next two years.
New York – February 23, 2010 – TheInfoPro, an independent research company for the IT industry, today released the results of its Information Security Study, which showed that 40 percent (40%) of enterprises are planning to increase their 2010 security budgets. …
Possibly Related Posts:
- Spending on Information Security Continues to Outpace the Rest of Corporate IT According to Latest Bi-Annual Study of the Global 2000 by TheInfoPro
- Enterprises Report that Internal Change & Learning are the Biggest Roadblocks in Moving to Cloud Based IT Infrastructures
- Latest IT Market Study From TheInfoPro: F1000 Enterprises 2011 Storage Spend Continues at a Strong Pace
- TheInfoPro Server Study: Operational Efficiencies, Not Compensation, Give Larger Organizations Cost Advantage Over Midsize Companies
- Fortune 1000 and MidSize Enterprise Organizations Say Immediate Spending Includes Telepresence and Unified Communications
Infrastructure vs. Applications: Pressure for Unified Global Delivery Services Causes Significant Changes in Infrastructure Spending
For several years TheInfoPro has tracked a move to the commoditization of certain infrastructure elements (e.g., servers, low-cost storage tiers). The economic crisis has made companies work hard on “optimizing” their infrastructure, as demonstrated by technologies related to this scoring so high on the TIP Technology Heat Index®, which gauges the immediacy of user need and planned spending.
So what can we take from this event as we triangulate the CIO’s comments with our data from interviews with decision-makers who represent more than $23 billion in buying power for IT infrastructure? Several threads become apparent:
1) Spending on Optimization: Investments in technologies that drive down the ratio of spending on IT infrastructure have dominated the top quartile of our respective Heat Indexes for the better part of the past 18 months and will continue into 2010. These include:
- WAN Optimization – Our guest CIO cited his company’s WAN being 50Mbps, and it is optimized to 400- 600Mbps.
- Thin Provisioning – This is enabling storage teams to increase their utilization rates dramatically, and to put off purchases.
- Deduplication – Well-documented success for backup environments. Look for more primary storage to leverage this in 2010.
- Server Virtualization – Nearly ubiquitous today for non-mission-critical applications, we are starting to see more discussion of production applications moving to virtual environments.
- Replication – Low-cost replication solutions are starting to minimize the use of expensive backup solutions.
- Unified Communications – Leveraging this (e.g., VoIP) continues to be deployed in an evolutionary fashion.
2) Cloud Computing: Internal vs. External? The discussion of external cloud usage has been a catalyst for the move to drop the ratio of infrastructure spending. Many organizations are able to show that the cost differential to move external is minimal. In addition, the ability to allocate more spending to the applications allows IT to be a better business partner and to help drive revenue-generating activities.
Caveat emptor – the shift to internal cloud computing architectures does not come without some pain, specifically around licensing costs. As software companies try to use the move to more processing power via internal cloud in a virtualized environment to mean more money for them, this is becoming a contentious point. This is one of the primary drivers for more things being considered to run on open source. Stay tuned for more work on this issue.
3) Open Source: Linux continues to grow in the enterprises mostly at the expense of Unix. One key reason is the software licensing issue mentioned previously; companies are trying to avoid additional licensing costs as more processing power gets shared by an application in a virtualized/cloud environment.
4) Standardization: From 2002 to 2008 we saw a lot of “flavor of the month”/“one of everything” mentality. This has changed dramatically and will continue to do so. Operating expenses have been decreased dramatically since fall 2008, and standardizing is a means to lowering spending. In addition, the aforementioned move to internal cloud is also a driver to move to standard offerings.
5) Procurement Strategies: In our work with enterprises, we see firsthand the sophisticated techniques being deployed to drive down the spending with their infrastructure providers. As an example, expect reverse auctions to procure servers and storage to become common by the end of 2010.
The biggest challenge we hear in our interviews with the Global 2000 is the ability to manage this new cloud-like infrastructure – as our research has shown for the better part of the last year, meeting SLAs, provisioning, lifecycle management – all come into play here and will be a key battleground for the vendors to provide an effective solution.
TheInfoPro has a research practice geared specifically to cloud computing and will continue the investigation of the storage, servers, networking and security markets. In 2010, we will report on the reality of the decreased spending ratio on infrastructure and how it is evolving. We suspect the trend toward changing the ratio of application to infrastructure spending will be a driver for more M&A activity by the technology providers who, not to sound cliché, are looking to become a “one-stop shop” and provide “one throat to choke.” More details on our M&A thoughts on this can be found in Barron’s October 2009 cover story: http://online.barrons.com/article/SB125633689630504703.html.
Possibly Related Posts:
- Unified Communications Solutions
- Heat Index Reveals Hot Infosec Technologies
- The Ascent of 10GigE
- Storage Vendors See Mixed Q4
- High-risk Staff? Executives and IT Are Equally Risky
RSS Feed