Data center Meltdown – Cause, Effect and Examples

October 21, 2014
Quincy Johnson

Author

Quincy Johnson

Digital Marketing Specialist at SoftwareONE | Linkedin

Maintaining a functioning data center is crucial for business.Not only is consumer interest for purchasing online is growing, but the medium in which they do so is expanding exponentially. For e-commerce sites, a data center meltdown could mean thousands of dollars per second, as was the case for Amazon when it lost an estimated $1,100 in net sales per second.

 Some staggering e-commerce statistics that further emphasize the need to maintain a healthy datacenter:

  • E-commerce grows 23% YoY, with 51% of American preferring to shop online in 2018.
  • Millennials are also increasingly depending on voice assistants for shopping, with 50% predicted to use such omni-platforms by 2020.
  • According to industry experts, 1:1 personalization will for a hyper-targeted shopping experience
  • In 2018 it is more likely that e-commerce customers are interacting via mobile phone, rather than desktop

 WHAT IS A DATACENTER MELTDOWN?

Meltdown made very simple is a vulnerability that can be found in many computer processors used for data centers. A meltdown can have various causes, the most common being malware attacking the system or overheating. Firstly, malware brings the entire system to break down by accessing protected areas of a chip, where confidential data such as passwords are stored. Secondly, a datacenter has numerous components, all of which can overheat. This is why it is important to maintain ideal data center conditions.

HOW TO PREVENT DATACENTER MELTDOWNS?

Most meltdowns can be avoided by keeping your software up to date, if your data center is of smaller size. Tech giants such as Amazon designed their own technology, which tells electrical switchgear how to act once the data center is experiencing a meltdown. It is mid-sized companies have the most difficulty preventing their data centers from a meltdown.

This prospect provokes the need for all of us – large or small – to question the health of our IT Infrastructures:

10 BIGGEST DATA CENTER MELTDOWNS AFFECTING E-COMMERCE

 

1. Intel Meltdown in Millions of Chips 

The most recent meltdown, is also the biggest one. Due to a flaw in Intel’s chip architecture, which is in almost every single processor sold since 1995, all those computers could be at risk of a potential meltdown. This includes all of Apple’s Mac and iOS devices ever produced. It was discovered just a few days later, in January 2018, that additionally to the meltdown threat, the CPU data cache timing within the same chips could be abused to abstract information from all affected devices. Currently, Intel is being sued by the City of Providence as well as by multiple companies. Various tech giants reportedly are looking at switching to rival products. A devastating drop in earnings is expected in Q1 2018.

2. Delta Airlines – $150 Million in a day

Thousands of passengers were grounded in the height of travel season in August 2016, as Delta’s data center was breaking down due to a power outage in Atlanta. As consequence, no planes could land or start, no passengers could be checked in and the communication system shut down – meaning no passenger could be alerted via email or their App.
Delta is not the only Airline battling data center issues, Southwest had to cancel flights for 3 days only two weeks prior and United experienced a meltdown in July 2016.

3. Amazon Cloud and popular related internet services 

As Amazon’s Cloud, their Simple Storage Software (S3) to be exact, went dark so did services relying on the service, such as Slack, Trello, Dropbox, Amazon Prime Musik, Giphy and Imgur to name a few. Many smaller businesses, heavily relying on a combination of those services, suffered majorly from this six hour breakdown. S3, promising to have 99.999999% durability, was brought down by a typo from an employee, who then removed a number of servers triggering a full restart of the Amazon North Virginia data center.

4. Amazon Web Services (AWS) – $66,240 per minute for 30 minutes

Although many of AWS’s IaaS customers were unaffected, Amazon.com itself suffered a major blow to its revenue stream as its e-commerce site went offline for nearly 30 minutes in 2013. However, Amazon stands as the symbol of versatility in the face of adversity, knowing data center outages are bound to happen and that Disaster Recovery is as essential to the cloud as the platform itself.

5. Google – 5 minutes: 40% global web traffic decrease

For the 5 minutes Google went MIA, the World Wide Web witnessed a 40% depletion in global web traffic. Although 5 minutes sounds like an inconsequential timeframe, $14.1B in reported Q2 2013 revenue resulted in a jaw-dropping $108,000 per minute loss.

6. Sears – $2.2M loss in profits, $2.8M to fix

This is an especially unfortunate case as third-party vendors typically promise a 99.9% uptime. With 4 power supplies infected at separate intervals, Sears suffered extensive downtime before resorting to generators to bring their servers back online. As if it couldn’t get any worse, the generator failed and Sears had to rent another for $13,500 per week – talk about insult to injury…

7.  Azure – Fortunate timing for a major release

Microsoft Azure experienced several outages throughout 2013, but dodged a potentially crippling blow on the eve of its Xbox One launch. Fortunately, Microsoft was able to correct the issue before consoles went on sale a few hours later, but other major Microsoft cloud services running on Azure such as Outlook and Office 365 went offline during the blackout. Pearl of wisdom – if your data center is going to crash, do it before a major release!

8.  New Jersey – The DMV just got worse…

New Jersey’s datacenter is no better off than your average SMB. Three times in Q3 2013, New Jersey’s IT systems shutdown due to minor issues such as power outage or fire alarms (without the ensuing fire), two issues which a separate, offsite datacenter can readily to go-live should the primary datacenter shutdown. 39 Motor Vehicle Commission offices across the state suffered a blackout, further distressing New Jersey citizens eager to renew their driver’s license or file tax information.

9.  Sony – Location, Location, Location

Just when we think we have all the technical ramifications considered, Mother Nature reasserts her dominance. The 2011 earthquake and tsunami devastated Sony’s 2011/2012 fiscal years, netting a total $3.18B loss. The earthquake measured a magnitude of 8.9, striking just off the eastern shore of Honshu Island; six Sony manufacturing facilities housing a variety of Sony technology were stationed nearby. Despite our best efforts to ensure the IT infrastructure’s technical standards, identifying the least hostile environment to house your organization’s IT backbone is essential to ensuring its structural integrity.

10.  Target – Pay attention to your security warnings…

And of course, the infamous Target debacle. Apparently various security systems including Symantec End Point Protection and FireEye – the software used by the CIA – caught the initial stages of this malware. Lesson learned – if your anti-malware system warns you of a security threat, don’t sing your own death knell by ignoring it.

Although we like to think IT systems are becoming increasingly infallible with each technological innovation, the fact remains that data centers are no more perfect than their creators. Monitor your datacenter’s health and implement a Disaster Recovery solution to prevent yourself from becoming the next data center meltdown headline.

Contact one of SoftwareONE's experts

 

Leave a Reply