Will the cloud lose its economic advantage?
Jim Kane warns about cloud computing suffering the same fate as 1970s era of time-sharing
Cloud-based solutions have established themselves as the clear choice for the production and distribution of information.
The underlying technologies generate powerful economies of scale in applications processing, and the relatively low cost of network-based distribution further enhances the economic advantages of cloud architectures.
However, today’s enablers could become tomorrow’s constraints and it’s not too early to ask how today’s cloud-based architectures can avoid becoming the ghosts of 1970s era time-sharing systems.
Every dominant computer services architecture - mainframes, time-sharing, PC-based, client-server, etc. - has eventually lost its economic advantage. The economics of any architecture is driven by trade-offs between two costs – processing and distribution. Cloud architectures rely on efficiencies from the concentration of processing capability and storage capacity, and the relatively low cost of network transport.
Cloud architectures are more likely to reach diminishing returns because of the network side of that equation rather than processing or storage costs. Therefore IT executives are well advised to put a new focus on understanding network cost and performance drivers to maximize the return on investment in cloud-based solutions.
Recent analyses on network performance, traffic distribution, and mobile technology suggest where IT executives might begin to focus. Performance integrity measures actual network performance vs. the stated capacity for which a customer is paying. The Federal Communication Commission’s 2011 and 2012 reports, Measuring Broadband America, quantify significant variation between actual vs. stated capacity among telecom providers. These results are based on sophisticated measurement and analysis methods, and have driven performance improvements.
While the FCC’s focus has been exclusively on households, comparable information for a business or agency becomes a basic management responsibility to ensure not only the viability of cloud architectures but also the validity of their costs.
Cloud architectures and “Big Data” are a recipe for network congestion. Comparison to highway congestion management systems provides perspective on how having management insights on the distribution of traffic and implementing appropriate controls can ensure that network performance meets users’ expectations, rather than the cyber equivalent of being stuck in traffic.
For example, consider the amount of traffic that cloud-based applications generate, and how data traffic is distributed among content, overhead, and what some call “clandestine” packets. This latter category is the most rapidly growing and frequently includes information being either collected from or distributed to a user to feed algorithms over which the user has no control.
An emerging management challenge is to understand the not-so-subtle impacts of how traffic is distributed and how applications drive that distribution and associated network performance and costs.
Mobile technology has obliterated the demarcation between business and personal time, and the associated costs of business vs. personal communications. Smart phones and cloud architectures enable workers to access ever more business applications . . . and generate ever more traffic and costs.
The Bureau of Labor Statistics recently reported that household spending on phone services rose more than 4 percent last year, driven largely by mobile data. The bring-your-own-device rage and the emergence of traffic-metered pricing plans create management pressure to understand the cost implications of continued growth in mobile computing and the impact of cloud-enabled applications upon that growth.
Being in a technology-driven business is like being the parent of an 11 year-old – what got you through the past three years won’t get you through the next three years. Concerns about the security, control and reliability of cloud solutions are yesterday’s battles. Tomorrow’s focus should be to understand the network side of cloud solutions to maintain the economic advantage of this powerful architecture.
Jim Kane is an advisor to companies in the federal market, and former president and CEO of the Systems and Software Consortium Inc.