Is the new question answering service from Wolfram Alpha a big stud or a big dud? According to a recent post from the Wolfram Alpha blog, we will all get to decide for ourselves when the service officially launches on May 18. Criticism aside, there is plenty of evidence from interviews and demos that Wolfram Alpha will be good at some tasks and bad at others, but overall a worthy addition to our information gathering toolboxes. (check out our previous coverage)
One of the most interesting things about the Wolfram Alpha launch is its hugely publicized nature. Most internet services launch quietly in an alpha or beta stage and then slowly improve features and access to the public as the kinks are worked out. Wolfram Alpha has forgone this proven technique, instead going for a much publicized “all at once” launch, creating an enviable level of interest, but also enormous expectations that may be impossible to meet.
The much publicized launch also creates a serious (and expensive!) engineering challenge for the Wolfram Alpha operations team responsible for building the infrastructure for the service. On May 18 Wolfram Alpha is going to receive an enormous influx of traffic as everyone and their dog attempts to give the service a go. Traffic on launch day will easily be an order of magnitude greater than what the service will normally see in its immediate lifetime afterward. This will require an expensive order of magnitude increase in the number of computers and networking gear needed to support the first day launch, significantly increasing the day one complexity and failure probability of the service.
4 Comments