5th Circuit and Server Efficiency
Much has been made, at least privately, of the 5th Circuit’s refusal to accept the “server efficiency” argument put forward by the University of Texas in the recent White Buffalo Ventures decision.
The answer to why the “server efficiency” argument was rejected is placed within the plain language of the decision. Notice:
We reject, however, the proposition that the ITC policy is no more extensive than necessary to secure the state’s second substantial interest, which is the efficiency of its servers.
One might persuasively present evidence that that spam, taken in its entirety, affects the
efficiency of email servers; indeed, that appears to be what UT has proffered; it submits a list of between 1,500 and 2,000 blocked IP addresses. Updegrove testified at the May 2003 Preliminary Injunction hearing that UT’s “system” would not be able to function without
these blocks. Such testimony is common where server efficiency is offered as a state or
private interest in Internet litigation.We must nonetheless consider the evidence in the light most favorable to the nonmovant. See Ford Motor Co., 264 F.3d at 498. Moreover, the challenged regulation should indicate that its proponent “carefully calculated the costs and benefits associated with the burden on speech imposed by its prohibition.” Cincinnati v. Discovery Network, Inc., 507 U.S. 410, 417 (1993) (internal citations and quotations omitted). (pg. 10)
The second paragraph there tells the tale: “We must nonetheless consider the evidence in the light most favorable to the nonmovant.”
When conducting a summary judgment hearing the Court is obliged to take all testimony at its weakest for the party trying to get the Court to dismiss the case and at its strongest for the other party. The way to get around this obligation is to provide the Court with information that backs up the allegation.
If the University of Texas, as the party trying to get White Buffalo’s lawsuit dismissed, wanted to truly rely upon the server efficiency argument, then they needed to present testimony supporting the argument. They tried to do this, but only by providing the trial court with the list of IP addresses they were blocking. There’s nothing wrong with doing this, but there appeared to be no link between this list and the efficiency of the servers other than the blanket statement made by Daniel Updegrove that if they couldn’t block that UT’s system would not be able to function.
The obvious question here is “why?” What is it that would happen if those blocks were removed? Would the server start billowing smoke? Would the hard drives rattle apart? Would the system become so slow as to be unusable?
But even more than that, there is the question of fit with White Buffalo’s spam specifically. In an effort to blunt the server efficiency argument, White Buffalo agreed to both rate and time limiting. This is certainly an offer that body-part-enhancement and drug spammers are going to refuse to make. Thus, White Buffalo is acknowledging the server efficiency argument and asking for some way to lessen that effect through an agreement.
If the University of Texas wants to continue with the server efficiency argument, they would have to find a way to tie in the use of lookup tables to ensure that mail is blocked except during the agreed-to hours, man hours in server configuration, and man hours in negotiations for rate and time limits (which would have to necessarily be someone with some ability to bind UT to a contract and a high degree of familiarity with the mail server and its configuration).
This is not to say that something couldn’t be done by UT in order to use this argument, but White Buffalo did a masterful job at taking this argument away from them at summary judgment by offering to work with them.
This is what the 5th Circuit means when it writes:
For the server efficiency rationale to pass muster under the fourth prong of Central Hudson, spam filters must block a set of spam that poses a legitimate threat to server efficiency.
This is not to say that UT need draw granular distinctions between types of spam where drawing them renders filtering economically infeasible. It, however, is to say that where UT may easily use certain types of filters—e.g., time of day and volume filters—UT should use them rather than categorically exclude all unsolicited commercial bulk email. If those types of filters are economically infeasible, that evidence should be in the summary judgment record. The current record reflects only that UT does not employ such filters because legal spammers are subjectively “misusing” the system, not because they are overburdening it. (pg. 11)
The problem with the server efficiency argument then, is that saying “the cumulative effect of all spam harms my server” does nothing to blunt the counter-argument of “we’ll work with you so that my spam does not harm your server.”
And now it’s time for the one, great, big, huge caveat to all of this: The University of Texas is a political subdivision of the State of Texas. The reason that these arguments hold any water at all is because a political subdivision of a state is not allowed, under normal circumstances, to prohibit commercial speech unless it can show that it has some rational reason for doing so.
You may do what you please with your own personal, private property. Even block White Buffalo Ventures.
MickC @ August 8, 2005