As usual for this time of year, we like to review some of the big stories of the last 12months and look at the impact on our industry. This year has been a bit of mixed bag with plenty more acquisitions, outages, cyber security and the odd racoon for good measure.
One fix, two fix, three fix, more
So, the year began with a double whammy of security flaws named Spectre and Meltdown. These flaws were present in the fundamental design of most processor chips, requiring a change to the way operating systems worked and this led to performance hits for some hardware platforms. Despite hot fixes being made available quickly, problems continued throughout the year as more vulnerabilities were discovered. Multiple bugs in IBMs Notes software products resulted in them dropping it like a dead donkey when HCL offered them $1.8Billion. On top of all of these troubles, chip makers announced price rises in February of around 20% which hit the data centre hard amid rising costs.Money, money, money
Following the trend from last year which saw growth in the data centre industry of around $20billion, cloud giants continued to pour billions into their growth programme's. Server Farm Realty were among a number of operators to open new data centre's in the UK, along with Chinese tech giant Alibaba who opened 2 new London based data centres. Further afield, Brookfield entered the market after it bought up AT&T's data centre estate and Digital Realty expanded their operations into South America with the acquisition of Ascenty. IBM made a game changing move to became the largest hybrid cloud provider by buying up REDHAT in October.However, it wasn't' all good news as Apples plans for a data centre near Galway in Ireland were scrapped due to planning delays and conservation issues. Quickbooks owner, Inuit, also fell on hard times after moving their core services to AWS forced them to sell their data centre's to H5.
Despite the rise of big cloud providers, a new study has found that on-premise data centres are thriving. The study by Market research firm IHS Markit, found that most enterprise data centre's expected to double the number of physical servers in the next 12monts. These enterprises are transforming their on-premises DC's to a cloud architecture citing key drivers for this as security, performance and scalability.
Ninety-Nine Problems
There were plenty of problems in 2018, kicking off in February with downtime at Yahoo preventing BT and Sky customers accessing email from mobile devices. Things were bad over at Capita, after 60 outages in 6 months, they were forced to move their services out of the own data centre with staff blaming skill shortages. In June an accidental discharge of the fire suppression system brought down a data centre in New Jersey, USA which took out local government sites for several hours. Another fire suppression discharge also took down a Microsoft data centre which caused widespread problems for Azure customers across northern Europe. Things didn't improve for Microsoft, as in September they had another Azure outage as a result of a 'cooling issues'. Our sources close to the tech giant informed us of another fire suppression discharge which shutdown the cooling systems that led to key equipment overheating and shutting down.Things were much worse for TSB as, after weeks of problems with the roll out of a new system, their head of IT at the UK based bank had to step down as customers took to social media to vent their fury after being unable to access their cash and pay bills. When October rolled around, it was Virgin Media's turn as a power failure in one of their data centre's caused days of downtime for their customers. And finally, in December, O2 dropped the ball when a digital certificate on telco hardware was not renewed and left customers unable to use their mobile phones for 48hours.
Who's afraid of a cyberattack?
New data protection regulations came into effect across Europe this year in the form of GDPR. There was chaos across the Internet as some websites simply blocked access from European users and despite the new regulations being designed to protect personal data, they actually enabled cybercriminals to hide their identities from law enforcement. In UK politics, it was slightly ironic that one of the first fines handed out under the new regulations was to the company handling data analytics for the Vote Leave campaign and across the pond in August, Microsoft thwarted cyberattacks on American political parties by grabbing dodgy domains.However, despite the American Government warning in July that hackers were turning their attention to supply chains as a method to get around cybersecurity measures, it didn't stop 30000 defence worker details being breached when one of their suppliers was hacked in October.
And finally...
Mother nature took another swipe at the industry again this year as July saw record breaking temperatures across the UK put data centres to the test. Meanwhile over in Seattle, USA a rogue racoon managed to black out a big chunk of the cities power grid for several hours. Hurricane season also caused havoc in September when Florence smashed in to Americas Data Centre rich east coast and Azure was once again in the firing line when a key data centre for Microsoft cloud in North America took several direct lightning strikes which took out their power and cooling systems.In august, The Uptime Institute published its annual data centre survey results and one of the major findings was that our industry is unprepared for climate change. Many facilities are not built to withstand the risks which extend beyond the data centre itself, to networks, collocation partners, utilities, infrastructure and suppliers. It found that even small changes in temperature for a certain number of days a year can cause problems for data centre operators. It is also becoming harder to recruit and train staff with the skills required to operate and support increasingly hybrid IT environments. There is a growing need for new skills as software defined and automation technologies are being deployed which may require less staff but they typically require skills that are not readily found in traditional staffing sources.
As region-wide disasters become more common, planning an organisation’s IT resiliency becomes more complicated and the skills shortage is driving up both costs and risks. Planning can no longer be isolated to equipment testing and emergency drilling procedures but instead must be conducted in the context of an organisation’s broader emergency and business-continuity plans.
So, it just leaves me to say thank you to all of our readers and wish all of you a happy new year.
You can follow me on twitter at https://twitter.com/datacentreman for the latest news and analysis on the data centre industry