Computing on Cloud Nine
By Dick Friedman
Google is trying to convince consumers to replace their personal computers with bare-bones PCs that do not contain software, store data or do any computing. A Google server would do all that. The bare-bones PC would be connected via the Net to a Google server, and would be used to enter data, display the result and, if desired, print the result. It's called "cloud computing," and if it works for consumers, it could eventually replace distributors' computers. Microsoft, IBM and others that license software are trying to determine how to provide cloud computing without killing the revenue stream from licensing.
Cloud computing is not a new idea. A previous article, "Distributors May Not Need Computers Anymore," described an early version of cloud computing, Software as a Service (SaaS). SaaS was pioneered in the business world, by salesforce.com, not Google, but Google is the first company to be in a position to make it a global reality. Others who tried to provide business SaaS found that potential users were indeed tempted by low-cost, almost worry-free computing, but frightened away by some serious concerns. Here is a look at one sketch of what could well be the future of business computing.
Is Cloud Computing Real or Blue Sky?
Definitions of cloud computing vary, but here is this author's definition: True cloud computing includes capabilities that IT people use to create or modify remotely stored software, build Web sites and perform other technical functions. Distributors do not need to know about the nitty gritty of these technical capabilities.
Distributors would not possess servers or PCs that now store data and software, and do the computing. Nor would distributors use fully loaded PCs that are now connected to the servers and function as terminals. Terminals would be "netbooks," Internet-capable cell phones and iPad-like tablet computers, none of which contain a hard disk drive or a means for loading software or data from a hard media (such as a DVD). A slimmed-down operating system, Internet browser and security software would be the only software resident on each of the terminals used for cloud computing; no application software (e.g., sales order entry) or data would be stored on them. One or more printers would still be needed. Each terminal would be connected to a Web site via a high-speed connection.
When someone using a PC as a terminal for cloud computing wanted to perform a business function such as keying in data from a vendor's invoice, that person would see and use the same type of menus and point-and-click methods that are used today. Different and smaller menus would be needed for cell phones and iPad-like devices – menus that are largely nonexistent today. The user would not necessarily know that all the keyed data was being stored on some distant computer (not necessarily in the USA), that all the software resided there, and that all the computing was taking place there. Ignoring the issue of PC size, cloud computing already exists; a person who sells on eBay is using eBay's software, not software on a PC, and storing data (e.g., pictures, descriptions, etc.) on eBay's computer.
When a cloud computing user needed help, that would be handled by someone employed by the operator of the Web site, and based in some remote location, not unlike what often happens today. If a "slim" PC, printer or data communications "box" malfunctioned, the "repair" would usually be fast and simple – unplug the sick one, and plug in a new one (which would automatically be recognized by the operating system at the Web site).
Unlike pre-year-2005 attempts at cloud computing, the reason it is likely to work now is the combination of almost-universal use of the Internet, huge advances in low-cost and mobile computing power and high-speed fiber optic cable that wires the world together.
Every Cloud Has a Silver Lining
Money. Distributors that replace their ERP systems with cloud computing would not have to pay hundreds of thousands of dollars for a software license, and then pay tens of thousands of dollars each year to get software upgrades and support. And distributors might not have to pay hundreds of thousands of dollars for installation assistance, education and training; do-it-yourself on-line education and training might be sufficient for some distributors. Costly servers and their expensive annual maintenance contracts would disappear, as would beefy PCs that are used as terminals. Furthermore, distributors would not need an IT person to handle questions and hardware problems, which would save dozens of thousands of dollars each year. A distributor could dramatically reduce IT expenses.
There's no such thing as free computing. Cloud computing is a pay-as-you-go business model that providers offer in two flavors. Plain vanilla involves a monthly fee for each registered user, with limits on how much activity is covered by that fee. Rainbow delight includes a fee for data conversion and training and education of users, and distributors would pay for the "resources" used (amount of data stored, amount of data processed, amount of data transmitted to the server and transmitted back for display/printing). Regardless of flavor, cloud computing would provide more resources than any distributor but the very, very largest could afford, at PC prices.
Upgrades and support. The concept of buying software upgrades, or paying for them via annual fees to software licensors, wouldn't exist. As soon as the service provider enhanced the software, all users would be using the same enhanced version, often without knowing that a change occurred.
Nor would a subscribing distributor have to worry about slow terminal response time; the service provider would monitor and forecast traffic speed, and would add processors/memory/dataciruits in advance of problems, maybe. Because most providers would use several interconnected computers that function as one, a failure of one computer would not impact subscribing distributors.
24/7 support for questions about software functions would be obtained via a problem reporting mechanism in the provider's software, and via phone support.
Every Cloud Has a Black Lining
Money. Unlike systems with licensed software, where costs are fixed, with cloud computing there is no cap on costs. Add users, and or do more, pay more. For a distributor, if additional users are registered or the volume of data or transactions increases, so will the cost.
Secure data? For distributors, having their data out there somewhere, but not on an in-house computer, is a discomforting thought. One way to get comfortable is to attach a backup device to one of the netbooks or thin PCs, and back up the key databases every night. Of course, that increases the cost.
Who would have access to the data? Who could see the transactions moving to and from the cloud? Couldn't hackers hack a cloud Web site?
One size fits all. Many distributors are able to get their software customized to handle unique business needs. In a cloud, what you see is what you get, and because it is meant for many distributors, getting changes made would be impossible. On a positive note, cloud software is specifically designed to handle numerous users and companies; earlier attempts at multi-company computing used software that had been designed for only one company at a time.
Need help? Take a number. Distributors using licensed software can call and/or send e-mail for help. Cloud computing should provide those mechanisms too, but nothing stops a profit-focused provider from discontinuing phone support. Try calling eBay with a problem.
How could it be this slow? Because the provider does not want to spend the money to add hardware or data communications circuits to overcome the slowness. Or because the Internet has slowed for reasons not under the control of the provider.
Chapter 11 or Chapter 7? If a cloud computing provider has trouble staying aloft, or folds, a 90-day notice of reduction in service or discontinuation of service may be a death warrant for the distributors who signed on. The data may have been backed up nightly, but there is no other software that can quickly and easily use that data. It can take at least 120 days to get a new system in place and the users trained to use it; it usually takes at least 180 days.
Dick Friedman is a recognized expert on warehouse operations, management and technology for fastener, tool, industrial and MRO distributors. As a Certified Management Consultant, he does not sell warehouse technology or systems. Based on more than 30 years of experience, he helps reduce warehouse mistakes and costs, often through inexpensive changes. Visit www.GenBusCon.com for more information or arrange for a FREE consultation; or call (847) 256-3260.
This article originally appeared in the Sept./Oct. 2011 issue of Industrial Supply magazine. Copyright 2011, Direct Business Media.
INDUSTRIAL SUPPLY MAGAZINE
The November/December issue of Industrial Supply magazine features an in-depth cover story about Kimball Midwest. Plus, we're featuring articles by contributing writers Dirk Beveridge, Frank Hurtte, and Mike Marks, as well as interviews with numerous product and distribution industry experts.
