Chaos for home work­ers af­ter Gmail crash

Five-hour out­age raises con­cerns that the UK is too ex­posed to a small num­ber of tech­nol­ogy gi­ants

The Daily Telegraph - Business - - Front Page - By Michael Cog­ley, Lucy Bur­ton and Si­mon Foy

LE­GIONS of home work­ers and anx­ious GCSE pupils were hit by a Gmail melt­down yes­ter­day, stok­ing fears that the coro­n­avirus boom in re­mote busi­ness has left Bri­tain danger­ously re­liant on a hand­ful of tech com­pa­nies. Google’s G Suite – which in­cludes the Gmail email ser­vice as well as Google Drive, Google Chat, Google Docs and Google Meet – was plagued with is­sues from around 6am in what ex­perts said was its worst ever out­age.

The tech ti­tan said it had re­stored ser­vices to users by 12.10pm. But in the mean­time the fail­ure wreaked havoc on em­ploy­ees try­ing to do their jobs re­motely, as well as ner­vous chil­dren forced to wait for their GCSE re­sults.

Data from tracker site Down­De­tec­tor showed the thou­sands of out­ages were con­cen­trated in the UK, Greece, Spain, Ger­many, France, Ja­pan and Malaysia. Those af­fected were un­able to log on, up­load files or send emails.

Derek McAu­ley, pro­fes­sor of dig­i­tal econ­omy at the Uni­ver­sity of Not­ting­ham, said the episode high­lights how de­pen­dent the coun­try has be­come on a small num­ber of tech busi­nesses that wield enor­mous power.

He said: “The is­sue is where every­one be­comes de­pen­dent (on a sin­gle ser­vice), a whole sec­tor can grind to a halt.

“But this doesn’t hap­pen when you have a more com­pet­i­tive and par­ti­tioned mar­ket­place.”

Gmail is the world’s big­gest email ser­vice, with around 1.8bn users.

Its G Suite ser­vices have surged in pop­u­lar­ity in lock­down with thou­sands of busi­nesses sign­ing up.

The com­pany is likely to play an in­creas­ingly vi­tal role in busi­ness life as em­ploy­ees aban­don the tra­di­tional nine-to-five. With nor­mal face to face com­mu­ni­ca­tion in the of­fice im­pos­si­ble, black­outs pose an ever greater risk.

The cause of the out­age is un­clear, but it was spec­u­lated that glitches in up­dates to the firm’s cloud ser­vice may have been to blame. It also co­in­cided with the first day back at school for mil­lions of US chil­dren, many of whom use Google prod­ucts to learn re­motely.

Tech­ni­cians from the com­pany are ex­pected to re­veal full de­tails of what went wrong in coming days.

There is no sug­ges­tion prob­lems were caused by a hack, but con­sul­tants said the fail­ure none the less high­lighted the risks if at­tack­ers are able to take down key email sys­tems.

Google apol­o­gised for the in­con­ve­nience and thanked users for their pa­tience. It said: “Sys­tem re­li­a­bil­ity is a top pri­or­ity at Google. We are mak­ing con­tin­u­ous im­prove­ments to make our sys­tems bet­ter.”

Jake Moore, a cy­ber­se­cu­rity spe­cial­ist at IT firm Eset, said there had been con­cerns about a Google fail­ure for many years.

He said: “As we all learn from mis­takes, it does el­e­vate a con­cern as to why this out­age was not able to with­stand such a force to knock it over and af­fect so many end users.”

If the internet is our in­for­ma­tion su­per­high­way, yes­ter­day’s mass out­age of Google ser­vices rep­re­sents the sud­den and to­tal clo­sure of the M25. Users up and down the coun­try who rely on the sys­tem for their liveli­hoods found them­selves con­fronted with the sim­ple Gmail mes­sage: “Oops, some­thing went wrong”. It was the dig­i­tal equiv­a­lent of the Road Closed sign.

Such is the pub­lic and pri­vate sec­tor’s de­pen­dence on soft­ware ser­vices pro­vided by Google and its ri­vals Ama­zon, Mi­crosoft and Alibaba that the five-hour out­age will likely be felt at GDP level. Never mind the frus­tra­tion felt by hun­dreds of thou­sands of home­work­ers, think of all the lost op­por­tu­ni­ties from meet­ings unat­tended, the lost con­fi­dence from work un­sent and the lost pro­duc­tiv­ity from re­duced out­put.

It all adds up: a tem­po­rary internet shut­down costs an ad­vanced econ­omy like Bri­tain’s $141m (£107m) per day, ac­cord­ing to a re­port from Deloitte and Face­book into the eco­nomic im­pact of dis­rup­tions to con­nec­tiv­ity. That’s equiv­a­lent to 1.9pc of daily GDP. A big hit, es­pe­cially in a re­ces­sion when com­pa­nies small and large are fight­ing for their lives and pub­lic ser­vices are stretched to the limit.

As a provider of big data soft­ware, I’m some­times asked if the UK’s re­liance on so-called cloud com­put­ing is a weak­ness. Ac­tu­ally, hav­ing ac­cess to enor­mous com­put­ing power at a frac­tion of the cost is a strength. It en­ables busi­nesses and or­gan­i­sa­tions to trans­form them­selves for the dig­i­tal age and scale up their prod­ucts and ser­vices. With­out cloud com­put­ing, it would be im­pos­si­ble to work at home dur­ing the pan­demic. Then our econ­omy re­ally would have fallen off a cliff. The prob­lem here is with re­silience. Ev­ery com­put­ing sys­tem has out­ages and they hap­pen all the time. Some are planned, usu­ally for week­ends, and some un­planned, usu­ally by ac­ci­dent. The is­sue is how long a sys­tem can be down – re­cov­ery time ob­jec­tive – and how much data can be lost – re­cov­ery point ob­jec­tive – with­out sig­nif­i­cant dam­age to the ap­pli­ca­tion, not to men­tion rep­u­ta­tion. Take Bri­tish Air­ways and its re­cent his­tory of IT fail­ures: the flag car­rier has suf­fered a series of out­ages caus­ing chaos and can­cel­la­tions for cus­tomers.

When we dealt in small quan­ti­ties of data, out­ages didn’t mat­ter so much. In the new world, we are gen­er­at­ing unimag­in­ably large amounts of data through wide­spread use of tech­nol­ogy in ev­ery as­pect of our lives. Un­for­tu­nately, big data also means big out­ages. When cloud providers say they can guar­an­tee against out­ages 99.99pc of the time, it might sound im­pres­sive but the cost of mar­ginal er­ror be­comes sub­stan­tial. Big tech firms could, and should, be guar­an­tee­ing dura­bil­ity of at least 99.99999pc.

The so­lu­tion to this prob­lem of leaky plumb­ing is what we call geo-re­dun­dancy. This means the abil­ity of ap­pli­ca­tions to “fail over” to ex­act repli­cas of data stored in en­tirely dif­fer­ent phys­i­cal lo­ca­tions. My com­pany, WANdisco, is pro­vid­ing these ser­vices to in­ter­na­tional air­lines, su­per­mar­kets and healthcare providers as they move core op­er­a­tional in­fra­struc­ture to the cloud.

Like it or not, the cloud is here to stay. It is big busi­ness. The mar­ket is es­ti­mated to be worth $371bn this year and is set to reach $832bn by 2025, ac­cord­ing to an­a­lysts. In the sec­ond quar­ter alone, the cloud gen­er­ated $3bn in rev­enue for Google, up from $2.7bn in the same pe­riod last year.

Com­ment­ing on the fig­ures, Sun­dar Pichai, chief ex­ec­u­tive of par­ent com­pany Al­pha­bet, said: “We’re work­ing to help peo­ple, busi­nesses and com­mu­ni­ties in these uncertain times.”

For Google’s 2bn-plus users, yes­ter­day was es­pe­cially uncertain. Ac­cord­ing to out­age tracker site Down­De­tec­tor, the prob­lems be­gan around 6am BST with wide­spread dis­rup­tion re­ported on Gmail, Google Drive, Google Chat, Google Docs and Google Meet. The out­ages were be­lieved to be con­cen­trated in the UK, Greece, Spain, Ger­many, France, Ja­pan and Malaysia. In a state­ment,

‘I’m asked if the UK re­liance on the cloud is a weak­ness. Ac­tu­ally, it is a strength. The is­sue here is re­silience’

Google apol­o­gised for the in­con­ve­nience and thanked users for their pa­tience and con­tin­ued sup­port. “Sys­tem re­li­a­bil­ity is a top pri­or­ity at Google. We are mak­ing con­tin­u­ous im­prove­ments to make our sys­tems bet­ter,” added the Sil­i­con Val­ley gi­ant.

The onset of coro­n­avirus has been a cat­a­lyst for change in the way we live and work. It has ac­cel­er­ated the adop­tion of tech­nol­ogy across the pub­lic and pri­vate sec­tors. Home work­ing is be­com­ing ac­cepted as a norm and pro­duc­tiv­ity is in­creas­ing as a re­sult, cer­tainly in the ex­pe­ri­ence of my com­pany and many busi­ness lead­ers I speak with. We all rely on the cloud to con­tinue to de­liver goods and ser­vices to our cus­tomers. As road users, we ac­cept the need for road­works and es­sen­tial main­te­nance. But we don’t ex­pect the sud­den and to­tal clo­sure of the en­tire mo­tor­way.

Google’s five-hour dis­rup­tion yes­ter­day was con­cen­trated in the UK, Greece, Spain, Ger­many, France, Ja­pan and Malaysia

Newspapers in English

Newspapers from UK

© PressReader. All rights reserved.