Free Lunch? Maybe …
Free online services are anything but.
Yet today, many of us are enjoying what seems to amount to a genuinely free lunch when we use online email systems such as Gmail and Hotmail, map and direction applications, and innumerable other online services available at no out-of-pocket cost to us. Unlike network television, the “price” for which is the time viewers spend watching commercials, many online services impose no comparable charge upon the viewers. For example, there is no obligation imposed upon the user to spend time viewing these website ads, and pop-up ads can usually be blocked.
What are the real costs of “free” online services?
These may include the costs of owning, leasing, or contracting for data handling, processing, transmission, and storage capacity. Then there are costs associated with personnel, real property, electricity, administration, O&M taxes, and a host of other expenses.
What long-term effects may we expect of the proliferation of free services that lead users to expect that these services should be available endlessly without costs to them?
Will this bonanza come to an end when the costs of ongoing expansion of data center processing power and capacity exceeds the appetite of advertisers to support such services?
Of equal or greater importance, will electric power be available to support continuing expansion of data center resources spurred by demand for free services?
Planning to ensure a sufficient baseload electric supply requires reasonably accurate projections of future demand in light of the nature of electricity (which requires that it be generated when used) and the lead time required for siting and building generation and transmission facilities.
Predicting future electric demand by the IT industry requires consideration as to whether the expected growth in user demand relating to free services (and other IT growth) will be wholly or partially offset by efficient cloud-based utilization and consolidation of data center resources and increasingly efficient equipment. In an earlier co-authored column (“The Cloud Changes Everything,” April, 2011), we pointed out that certain of the presumed efficiency gains expected to be achieved by the cloud are likely to be partially, or even wholly, offset by far greater usage by endusers facilitated by cloud resources and the energy requirements of transmitting information. Cloud growth is necessarily constrained by the availability of electricity supplies to power the increase in data center availability and capacity required to meet such growth.
Although increasing energy efficiencies are being achieved, as discussed in “The Pending U.S. Energy ‘Cliff’ and Opportunity for Industry” by John Tuccillo in Mission Critical’s January/February 2013 issue, such efforts may prove insufficient to protect the IT industry against future electricity generation shortfalls. For that reason, the IT industry may wish to take a page from the electric industry’s example of demand response in order to help stave off a situation in which meeting the demand for services requires data center facilities beyond those which can be supplied by the then available electric power.
At the risk of raining on a very large parade, we question whether the principles of electric distribution demand response can and should be applied to free internet services. In an electric utility franchise territory, the cost of providing electric service is allocated among the customers based upon the rate design approved by regulators. As California discovered in its initial experiment with electric utility deregulation, serious shortages can occur when utilities are required to charge customers less than the utility’s cost for electricity. The reason is simple — without proper economic signals, customers have little incentive to conserve. To make matters worse, increased use by utility customers, unconstrained by having to pay the greater true cost of electric power, can lead to cascading negative economic effects in addition to shortages (e.g., driving up the wholesale costs of electricity and fuel, causing inefficient plants to be dispatched, and overburdening transmission lines).
Similar electric industry problems occur when the cost of electricity serving apartments or businesses is included in a tenant’s lease. The absence of a cost-based incentive to conserve can lead to overuse such as air conditioning, other appliances, and lights being left on 24/7 regardless of whether the tenant is present. This is similar to the effect of free online services but exacerbated by the fact that there is no charge whatsoever for such services.
When a submetering proceeding was initiated by the New York State Public Service Commission over 30 years ago, the commission recognized that the responsibility for the cost of usage should be relegated to the end user when it stated, “[T]he hand which turns the switch should hold the pen that writes the check for the electric bill.” Similarly, the Federal Energy Regulatory Commission in August of 2006 issued its Assessment of Demand Response and Advanced Metering, which defined demand response as, “Changes in electric usage by end-use customers from their normal consumption patterns in response to changes in the price of electricity over time … to induce lower electricity use at times of high wholesale market prices or when system reliability is jeopardized.”
Applying electric demand response to no-cost online services meets an immediate hurdle. Since these services are provided free of charge, endusers experience no expense based upon the amount of usage, which precludes the possibility of demand response. Free internet services appear to be expanding, and this expansion (combined with other internet usage) is likely to bring about effects with which we are concerned — namely, the growth of data coupled with a greater need for data center processing and data capacity — and increasing use of electricity. We can expect ad-hoc responses by providers such as initiating charges for previously free online email systems and other free online applications based upon use. Such responses may be insufficient, since users may simply seek free services and precipitate a “race to the bottom” wherein it is difficult to equitably allocate costs without the adoption of some form of industry protocol or even governmental intervention.
Absent fees, there are certain downsides to free services, which might have the effect of moderating usage. Just as the enduser is not held accountable for the cost of services (at least not yet), it is likely to be very difficult to hold a provider of free services legally accountable to users. In a data center industry that measures itself by its uptime performance and strives to achieve continuous performance, going down for a period of time can create an uproar. However, a user of these free services is likely to be entitled to virtually no rights when something goes wrong — service suspended for a period of time, emails undelivered, wrong directions given, or storage vanishing. In many cases, not only is there no right against the provider, but a free service user will have waived his rights with respect to the provider under the terms of a “click on” agreement, which disclaims the provider’s liability.
Even in the case of a business for which an online failure of a free service might cause the business to suffer a loss or encounter liability, there is likely to be no remedy for the user against the provider. This is a case of “users beware.” Users should also be aware that, although, at a glance, there appears to be no disclaimer of the site’s liability, a disclaimer may appear in small and faint print at the bottom of the screen. That faint print may not even appear upon all screens on the site.
If there actually is no click-on agreement or no disclaimer, a user-plaintiff would have possible recourse against the provider only on the basis of negligence, if at all. In a lawsuit, the general rule is that a defendant is not liable in negligence, even if that person or entity did not act with reasonable care, if no duty is owed to the plaintiff. The classic law school case for this general principle is New York’s Palsgraf v. Long Island Railroad Co. from 1928. In that case, Ms. Palsgraf was standing at one end of a railroad platform when railway employees helped a man rushing for a train to board with a package that happened to contain fireworks. The package dropped and exploded. At the other end of the platform, scales fell on and injured Ms. Palsgraf. She sued the railroad and lost. The New York Court of Appeals held that the employees could not have reasonably foreseen the danger to her when they tried to help the man board the train and that, therefore, the railroad owed her no duty of care that could be breached.
For these reasons, it appears that making use of free services is at the user’s risk — but on the upside, it’s as close to a free lunch as you can get these days. As to the long-term consequences of free services, despite commendable and essential progress in energy and equipment use efficiencies, we foresee continuing and accelerating growth in the need for data center capacity giving rise to an increasing need for electric power. As such, it is possible that some form of demand response can help to mitigate that growth as it does in the electric industry.