Acquisitions
Skyrocketing growth in cloud services and outsourcing of enterprise data centres to colocation providers has translated into a boom for the data centre industry over recent years. The market has been in a state of continuous change throughout 2017 with pressure from customers' quickening the move to cloud, as well as data centre vendors chasing market share.Although the global market took a bit of a downward turn last year, the UK market is still strong despite Brexit fears. The last 12 months or so has seen many of the major cloud players take up residence in the UK, with the likes of Google, Microsoft and IBM all opening infrastructure regions in London. Looking across the industry in general, the UK is booming and London is one of four major European data centre markets currently seeing “unprecedented” colocation growth. Further good news for the UK market came in the form of British data centre developer Kao Data as they secured £33 million ($41.5m) in funding for the development of a large data centre campus located just outside London. The campus will become part of the upcoming science and technology park, constructed on the site where Sir Charles Kao first developed fibre optic communications in the 1960s. The first facility, Kao Data London One, is expected to open at the end of 2017. In addition, Newcastle can now lay claim to hosting the UK’s largest purpose-built data centre after operator Stellium decided to bring more server-based services to the North.
Read More
The market consolidation we have seen in recent years, showed no sign of slowing as the major data centre vendors have all been expanding by building and acquiring data centre facilities and the year kicked off with a flood of acquisitions. CyrusOne, QTS Realty, Equinix and Digital Realty all kicked 2017 off with strategic acquisitions. Additionally, privately-held Digital Bridge Holdings continued a fast-paced acquisition program by purchasing facilities from 365 Data Centres. Interexion joined the acquisition scrum in February when they purchased Vancis Holdings BV, including its data centres in Amsterdam’s Science Park, a major interconnection point in the market, where Equinix, their biggest competitor, has a large presence.
Equinix continued its programme of acquisitions by purchasing IO UK's data centres based in Slough. These data centres will be tethered to the other facilities on the Equinix Slough campus to add additional interconnection and capacity for their customers. These UK acquisitions can be viewed as a vote of confidence for London remaining a major global financial centre despite UK's Brexit vote. Equinix has also struck a deal to acquire Itconic, a Spanish data centre services provider. The deal will give Equinix new enterprise data centre customers and expands its footprint to new markets, where its clients can now extend the edges of their networks to and tap into submarine cables that will carry their traffic to and from Africa. Equinix ended the year by announcing another blockbuster deal, the $3.6 billion Verizon Americas data centre portfolio carve-out.
However, it wasn't all good news for Equinix. They were left with a lot of empty data centre space after LinkedIn consolidated much of its data centre footprint into large wholesale facilities. Some 1300 cabinets worth of equipment was moved out of Equinix data centres and left the colo scrambling to fill a $6.8million hole in their revenues.
Read More
The data centre wars heated up in June when Digital Realty acquired DuPont Fabros. Both companies have benefited from the surge in demand for cloud services and the California-based colo has been acquiring companies to boost their growth. The Digital Realty Chief Executive, William Stein, said the combined company would be the “home to the cloud”. In October, Digital Realty set up a joint venture with Japanese trading house, Mitsubishi Corp, to build 10 data centres in Japan by 2022. Tokyo-based Mitsubishi expects the data centres to help meet growing demand for information storage from customers of Digital Realty.
Read More
Rackspace joined the battle for market share in September when they acquired Datapipe, one of its largest competitors in the managed public and private cloud services business. Once the acquisition is complete, Rackspace will get new data centres and offices in markets where it currently doesn’t have much of a presence, including the West Coast, Brazil, China and Russia. In China, Datapipe also currently offers managed services on the fast-growing Alibaba Cloud.
Read More
Its not all good news in the data centre industry though. A number of colocation vendors have dropped out of the data centre business this year. Daisy announced in October that they were closing the doors on their own data centres and switching customers to Microsoft Azure Cloud. Eduserve followed shortly after, giving their customers the choice to either move to the cloud or move out.
Read More
Cyberattacks
Cyberattacks were big news this year and we kicked off with Barts Health Trust, the largest NHS trust in England suffering a major cyberattack in January. More problems for public sector followed when Aberdeen City Councils website was taken over for almost three hours in response to Donald Trump’s controversial travel ban on Muslims. Both organisations were quick to reassure the public that no data had been stolen.Big news came in March with the biggest ever leak of secret CIA documents. The whistle-blower website, WikiLeaks, published thousands of documents mainly focused on techniques for hacking and surveillance. The leak, named “Vault 7”, again raised questions about the inability of US spy agencies to protect secret documents in the digital age. Later that month, hundreds of Cisco customers were left reeling after it was revealed more than 300 models of switches were vulnerable to an exploit included in the trove of alleged CIA hacking tools. This leak would continue to have repercussions throughout the rest of the year.
Read More
One of the biggest cyberattacks of 2017 came in May, when the NHS and other public services were hit hard by the WannaCry outbreak. Hospitals and GP surgeries across the country were affected when the outbreak encrypted data on infected computers and then demanded ransom payment in bitcoin. Although no patient data was compromised, NHS staff were unable to access systems which resulted in cancelled appointments and operations. It later emerged that the NHS was not the target but it was part of a wider attack across the globe. The attack propagated through the use of the EternalBlue exploit which had been released in the CIA Vault 7 leak earlier in the year. Fortunately, the attack was stopped within a few days by emergency patches from Microsoft and the discovery of a kill switch that prevented the spread of the worm. The attack was estimated to have affected over 300,000 machines across 150 countries and it is believed to have originated from North Korea.
This attack was a wakeup call for many organisations but not everyone was a victim as security firms saw their share prices soar in the wake of the attack.
Read More
Downtime
One thing that was certain this year was that nobody was safe from the risk of human error and there were more than a few major outages in 2017.After shutting down ones of its own data centre last year following a failure, SSP's cloud broking service has again been dogged by availability problems after facilities engineers at an Equinix owned data centre caused a loss of power. The cause was apparently a “direct result of the ‘routine maintenance’ carried out by the facility engineers,” according to SSP, which says it ”was assured in advance that the ‘routine maintenance’ would not affect any services.”
Read More
In February, Amazon was in the firing line when human error caused an outage on the Amazon Simple Storage Service (S3) and it took down half of the internet with it. The incident caused massive disruptions. Big brand companies like Adobe, Expedia, Coursera, Medium and more were affected by the outage. Even the U.S. Securities and Exchange Commission (SEC) was knocked offline. The cause was identified as a typo in a command entered by an Amazon employee that caused a cascade of failures.
Read More
In August, it was Googles turn on the roulette of human error, when fat fingers sent the whole of Japans internet dark for a couple of hours but it was so severe that Japans government had some serious questions for internet carriers. The cause was finally identified as an incorrect border gateway protocol advertisement that blackholed all Japanese internet traffic.
Read More
In October, Microsoft fell victim to human error when routine maintenance on their fire suppression systems in one of their northern European data centres resulted in an 'unexpected' release. This caused the air handling units, responsible for keeping everything cool, to shut down and the ambient temperatures rose above normal conditions. The air conditioning was back up within 35 minutes and temperatures where back to normal shortly thereafter. However, some of the systems which overheated had shut down in an uncontrolled manner and so took longer to bring them back online. The disruption, which included service failures and latency issues for European Azure customers, lasted almost 7 hours.
Read More
As always, I have saved the biggest story of the year for last.
Airlines had it bad last year, with many high-profile outages, including the costly Delta Airlines incident. This year began ominously with an outage in the baggage system at Gatwick in May which created chaos and left passengers flying without their luggage.
A few days later, it hit the fan big time when all of British Airways systems went down for 2 days. The incident coincided with the first weekend of the school half-term holiday in the UK, one of the busiest periods for UK airports. There were chaotic scenes across the globe as hundreds of flights were cancelled affecting over 75,000 customers. Passengers were left furious as BA chief exec, Alex Cruz, took to twitter to apologise.
There was some initial finger pointing to outsourcing of IT services and rumours of a cyberattack, but BA dismissed these saying the cause was a 'power supply issue' in one of their data centres. A final statement blamed a UPS engineer who inadvertently disconnected a power supply. The power was then restored in an unplanned and uncontrolled fashion which caused physical damage to the systems and significantly exacerbated the problem.
However, with over 500 data cabinets spread across 6 data halls in 2 separate facilities, a single UPS failure should not cause this kind of problem. We all know that the critical infrastructure in a data centre, like any other IT system, can fail and this is why we build resilient systems. If a UPS was the root cause of the issues it still should not have affected both sites, therefore, I suspect that a lack of failover testing of the data centre and IT systems are the real reason the outage was so bad.
The outage is expected to cost BA in excess of £80 million but Alex Cruz has resisted calls for his resignation.
Read More
And just in case they thought they got away without anyone noticing. On the same day as the BA outage, a Capita data centre in Kent, which services the back-office systems for a lot of NHS trusts in the UK, also went down. The outage lasted 36 hours and was also attributed to a 'power failure'. Although the exact details of the cause have not been made public, I suspect this could also be chalked up to human error.
Read More
And Squirrels
With cyberattacks and outages making big news across the data centre industry, a more worrying threat has now appeared. In a bizarre series of events, squirrels and other small animals have carried out hundreds of attacks against the electrical grid causing data centre outages across the globe. They are responsible for more damage than any nation state as they chew through cables, climb into transformer boxes and store their nuts in the power distribution panels. The number of attacks against physical infrastructure by squirrels in the last year totals 1119, while humans account for just 3. It’s not clear yet what the industry is going to do about this new emerging threat but I am sure that there are 'top people' working on it.Read More
See you next year
It has been a very busy year for the data centre industry and we have seen increased threats from cybercrime and human error keeping everyone on their toes. Next year is likely to be even busier as cloud services continue to spread and GDPR comes into force. Unless these threats are properly addressed, it could be a very expensive year for many organisations.Read More
Last year, I warned of the dangers of not learning the lessons from previous failures and it appears that this year we, as an industry, still have not learned those lessons. Hopefully, next year, we won't make the same mistakes.
Read More
So, it just leaves me to say thank you to all of our readers and wish all of you a happy new year.
For the latest news and analysis on the data centre industry, follow me on twitter; @datacentreman