Q&A: The True Cost of the Cloud
Cloud computing is attractive for cutting IT costs -- but only if you've properly calculated the true costs in your ROI calculation. To learn how to better estimate cloud's true costs and maximize the benefits of cloud's scalability we spoke with Ray Solnik, president of Appnomic Systems, Inc., a company that specializes in predictive, application-centric, and autonomic management of IT infrastructure.
Enterprise Strategies: Budgets are tight -- how can IT accurately estimate the cloud's true, complete cost? What components are overlooked?
Ray Solnik: The two most overlooked cloud expenses are "people" costs and ongoing costs of leaving cloud resources up (versus leveraging inherent cloud elasticity to reduce resources when they are not required).
IT staff often emphasize compute and storage capacity, networking, memory and bandwidth because that is what we know. However, the Capex vs. Opex debate is typically not adequately assessed. Too many finance and IT professionals are still stuck on Capex and not the real costs of running assets or managing cloud vendors.
Another recent "Ah ha!" for many CIOs is the cloud service invoice. Some are surprised at how many services are spun up and left running and surprised at the resulting high cost. Many are beginning to realize cloud resources left on are significantly more expensive than managing resources themselves -- even if they are calculating people costs.
Some decision makers are actively screening cloud use cases and providing in-house, hybrid solutions for more sustained IT resource needs and are being realistic about management requirements. They are maximizing cloud business benefits while achieving optimal cost structures.
Given these factors, I wonder if many organizations are undertaking cloud initiatives that return a lower ROI than expected. What's your experience?
The ROI equation encompasses two key elements -- cost and revenue.
Recently, I am finding more IT VPs and directors, not just CIOs, are starting to understand and better deliver revenue growth to achieve ROI versus primarily cost management.
Interestingly, CIO.com recently published an article that 54 percent of CIOs aspire to general management positions, yet only 4 percent of CEOs come from IT. Revenue growth-oriented ROI projects are now a requirement for these aspiring future business leaders.
On the cost side, constantly up and running cloud infrastructure returns lesser ROI and, often, less than if internally implemented.
Over time, prices will continue to drop and the scalability benefits will far outweigh the value of in-house data center operations, but we're not there yet. Many software businesses, like ours (Appnomic Systems), are also enabling the up and down cloud resource provisioning to optimize cost savings.
Still, in many cases, nine months of cloud server fees could fund a self-owned and managed machine. Organizations maximizing ROI may internally operate their cloud, strategically optimizing hybrid practices, and manage environments via SaaS providers.
Ultimately, IT teams helping business partners crush topline growth opportunities via cloud apps and capabilities deliver the highest ROI. Cloud cost can be managed and optimized for adequate growth and revenue. Cloud-enabled revenue growth is where our partners, colleagues, and customers maximize ROI.
A colorful recent example of growth-driven ROI is the integration of an online pizza restaurant point of sale (POS) system with the pizza oven that took four minutes to heat up when an order came in. Previously, the process to take an order and deliver it took 30 minutes of which four minutes was often eaten by the oven warm up Post integration, the oven reheating began as soon as a POS order entry began. A huge portion of those four minutes were eliminated -- increasing cycle time by over 10 percent. That means fewer order "walk aways," fewer cold pizzas resulting in lost customers, and better customer experience.
Scalability is a key benefit touted by cloud proponents. Is this a checkbox feature that enterprises don't really use? Are popular service providers as infinitely (and easily) scalable as we believe?
Every enterprise I know considers scalability and their service providers' scaling approach. Google, Facebook, Twitter, LinkedIn, Netflix, and others have proven that massive and dynamic scalability is definitely achievable.
For those who do not have the scale of these companies, one way to assure scalability is to dual- or multi-source and deploy measurement systems across service providers that provide a single view of comparative performance across vendors. You can also require cloud vendors to provide more transparency and earlier warnings of potential issues.
The Open Data Center Alliance, ODCA, is one industry organization advocating more transparent standard measurement units so enterprise IT has a common view of various service providers. ODCA sponsors proofs of concepts to illustrate how to ensure cloud applications are truly scalable.
How can an organization determine how and when it will need to scale up or down? What are the bottom-line financial and service-quality ramifications of investing in a poor scaling strategy?
Monitoring and managing scalability at device levels will not survive long as the industry refines solutions to measure and assure delivery experience in the application layer with real user experiences in cloud applications and application stacks get more complex.
Enormous financial gains are achievable by applying advanced analytics to IT and through automated early warning alerts that help prevent rather than repair performance issues, enabling scalability. New pattern-matching technologies are also enabling applications operations staff to better have the application scale up and down based on user experience. There are now solutions that analyze patterns of how users interact with underlying environments and database usage -- know how and when to scale by applying predictive business analytics to IT environments. An "ounce of prevention" could determine scaling or not scaling, success or failure.
Cloud success belongs to those investing in architecture, scalable infrastructures, and innovative companies driving these technologies including Splunk, Appnomic Systems, Optier, and others.
There's some confusion about the best ways to get started with cloud initiatives. Which technologies/approaches are worth looking into and which are just hype?
Start by applying cloud initiatives in proven SaaS business categories such as CRM, sales and marketing automation, or workforce management to realize demonstrable business gains, and take incremental steps expanding to other applications. To mitigate risk, IT operations should test a public cloud environment for applications that are not proprietary, regulated, or confidential. There is no need -- nor is it smart -- to tackle an enterprise-wide implementation out of the gate.
The best approach: learn to crawl, walk, then run. It means fewer skinned knees or massive tumbles.
What are the biggest mistakes enterprises make in cloud initiatives? Other than accurate costing/budgeting and ignoring hype, what best practices can help overcome these mistakes?
I believe the number one mistake is betting on a single provider -- putting all your eggs in one basket with unproven companies or those spread too thin in areas outside their core expertise. Focus on specific problems, a focused vendor, and extend incrementally. Of course, for some applications, you will only need one vendor and, in that case, call at least three references.
What types of companies are moving most rapidly to cloud-based operations? What are the different ways an organization can implement its cloud strategy and how are they choosing to implement -- in-house deployment and management or a combination of application provider and service provider?
Most enterprises take a hybrid approach with a strong bias to quick cloud adoption -- across industries and geographies. Of course, some pockets are adopting faster than others.
What factors hold companies back from putting their toe in the cloud? How should an organization think about adopting/migrating to cloud?
One word: Fear.
Game-changing approaches are out there. Companies need not fear -- be smart, thoughtful on mitigating risk, and always have a back-up plan. Even when implementation is underway, scan the landscape for new opportunities. Invest and partner with your vendors to win.
What products or services does Appnomic offer relative to the cloud?
Appnomic Systems provides preventive IT performance management solutions enabling businesses to better leverage their IT operations for growth, profitability, and competitive advantage. Appnomic's AppsOne and OpsOne software leverage advanced analytics to automate both complex application performance management tasks and labor-intensive IT processes and data center operations. Appnomic customers are applying AppsOne and OpsOne advanced IT operations analytics and IT process automation technologies to prevent major IT incidents, accelerate break-fix system repairs, achieve new levels of compliance, and avoid major capital spends by better understanding application behavior complexities. The company's solutions are available as software-as-a-service (SaaS), enterprise licensed software, and as managed services from Appnomic or its partners.
James E. Powell is the former editorial director of Enterprise Strategies (esj.com).