WhatsApp, which is said to have nearly a billion users, went down in some regions of the world on Wednesday but there has been relatively little noise about it.
This, Andrew Martin, vice-president of Zerto in the Asia Pacific and Japan, told iTWire could be due to the fact that it is an app that mostly is used for non-business purposes.
Additionally, he said, it may be due to the fact that people who use WhatsApp have other messaging apps that they utilise and thus may not have missed it to the extent that they would were it just a lone app used for one purpose.
"They are not as inconvenienced, hence the degree of reaction is different," he said.
{loadposition sam08}But, Martin (below, right) hastened to add, the outage had highlighted once again the fact that, even though WhatsApp was a free service, its users will not accept downtime.
He agreed that this pointed to the fact that people now treated Internet connectivity in the same way that they treated gas, electricity and water – it should always be on tap.
WhatsApp, which is owned by Facebook, has not made any comment so on the cause of the problem, but tweets resulted in it being the top trending topic in the UK.
Many users were unable to send or receive messages.
Web service monitoring site Down Detector reported that issues with the app started at around 9pm in the UK (noon Thursday AEST). Problems were also reported in the US, South America and Asia.
Martin, whose employer is a company that offers software solutions for disaster recovery, said he could only offer educated speculation as to what had gone wrong.
He said that the issue could be software-related as features were being continuously added to the app.
"I heard they are testing a way to pin a message of one's choice to the top of the app (the same way that one can with Twitter)," Martin said.
While rolling out such features, there was always the chance that something would go wrong. That could be the reason for the outage being only in some areas, as each area could potentially be served by a different set of servers and such changes would not be rolled out en masse, but rather region by region, he added.
Martin said using systems which allowed for test failover — where one could test changes in a cordoned off area that was not live — and then rollback to an earlier working configuration was one way of avoiding such outages.