Amazon, the Media, and the Future of Cloud
Since cloud is a hodgepodge of technologies and internal and external services, trying to build a cohesive strategy, to optimize cloud can lead to a lot of circular motion. But one way to avoid this is NOT to put cloud first as that mysterious (and actually fictitious) "endgame;" as in the journey to the cloud, but to focus on key business objectives for IT and then see how and where cloud fits best. As Ive said more than once, the journey to the cloud from a purely logical perspective makes no more sense than the journey to VLANs.
But cloud has many parents beyond technology itself and these include aggressive marketing campaigns by the likes of Amazon EC2 and other service providers and vendors selling infrastructure and related products and services. It should be pointed out that many of those service providers are simply repackaging hosted services provided under other names, but who have adopted cloud for obvious marketing reasons. After all, if a service is delivered over a network, and can be flexibly provisioned (which can mean many things) and extended to on-demand needs (which can mean many things) and accounted for based on usage (which can also mean many things) you get to call it cloud (and be fairly faithful to the National Institute for Standards and Technology (NIST) definition).
But even more than marketing, the media has itself largely been the creator of cloud. Clouds prevalence in our minds would be nowhere near what it is without inflated media attention (a.k.a, hype), which, yes, I realize that Im contributing to right here. You can view it as an expression of age and its accompanying moral fatigue that I feel absolutely no shame in doing this.
However, as everyone knows or should know by now, the media thrives on creating trends and then destroying them, much in the way that demented children in horror flicks like to tear up their dolls and other intimates. Media does this, ostensibly, because it sells, which logically means that we, the reading public, are the true demented children here.
And so cloud (internal/external, SaaS, PaaS, IaaS, etc.) is finally getting to the point where the media should be hungry for pins, if not actual machetes, to stick in Cloud Barbie" and, if not render her headless, at least remove a finger or two. Given that, the media Ive read to date regarding recent Amazon's EC2 outage seems pretty responsible. But not so much that the over-arching specter of a demise in cloud preeminence isnt looming, or at least present enough to sustain the impression of drama.
- First of all, for a few of the more salient "facts" up to the date of writing this column as of April 26th:
- On Thursday, the 21st of April, at 5:16 a.m., Amazon's Service Health dashboard reported connectivity problems impacting its Relational Database Service affecting a broad area especially along the U.S. East Coast.
- This disabled some fairly popular websites such as Foursquare, HootSuite, Quora and Reddit, all of which are back up at the time of writing this column.
- At 10:35 a.m. on Sunday, April 24, Amazon reported that "We're in the process of contacting a limited number of customers who have EBS" (Elastic Block Storage) "volumes that have not yet recovered and will continue to work hard on restoring these remaining volumes."
- On Monday, April 25, Amazon reported that engineers were still working on issues surrounding its EBS.
- Amazon is planning to provide a detailed "post-mortem" of the root cause surrounding its outage and that its workers are "digging deeply" into the event.
These specifics, obtained through fairly consistent technical media coverage at least, are in large part characteristic of the coverage itself. However tones have ranged beyond this to headlines signaling that Amazon got a black eye from the outage (reasonable enough) to Amazons cloud nightmare to who gets the blame? Mass media, or at least the New York Times , have been reasonably responsible; as Steve Lohr's headline read, Amazons Trouble Raises Cloud Computing Doubts. It should be noted that in most cases, real opinions came from analysts like myself (though not from me personally until now, of course).
The articles I looked at also contained some good advice, such as:
- Any system that concentrates too much critical data in one place becomes vulnerable.
- If using cloud, design around it, just as you would back up any external critical service such as failover to internal resources
- Leverage a variety of cloud service options from different providers both to assess relative quality and cost, and to minimize impact of outages and degraded performance (my words summarizing several sources).
EMA and other research also indicates that skepticism towards cloud wasnt just waiting for Amazons conspicuous outage to happen. It was there before hand; due to both native IT skepticism and lessons learned from early cloud adoptions.
For instance, 70% of cloud deployments have required redoing or rethinking (EMA data) and Compuware data indicates that North American companies estimate organizational/business losses of about $1 million a year from degraded performance from cloud-based applications. The estimate is somewhat lower in Europe at $775,000. Dialogs with deployments have underscored a number of obvious risks, including security risks, e.g. Who knows what Google thinks is a violation of privacy? when it comes to managing data.
It should be stressed that performance issues are far more pervasive and harder to gauge than absolute outages. Amazons misfortune has, as is typical, been dramatized in the media by the obvious. But the future of cloud remains no fundamentally different post-Amazon-outage than it did before. It is neither salvation nor endgame but an array of new technologies and services that, ironically, as they mature and become more effectively assimilated, will lose in media visibility as the media moves on to something newer and more controversial.