How Short Outages Become Long Outages

Early in the morning of Friday July 27, 2012, Hosting.com experienced an 11-minute outage. Although service was restored very quickly, many customers weren't prepared and experienced hours of downtime as a result (source).

The key story here is that even though a few minutes of hosting provider downtime is probably well within the parameters of the service level agreement (SLA), the customer's actual downtime far exceeds that. I'm going to quote my own blog from just a couple weeks ago because it accurately sums up the situation:

Your cloud (or other hosting) provider no doubt promises a certain amount of uptime in their service level agreement. Let's imagine that allows one hour of downtime per year. If they have one minor problem it could cause downtime of just a few minutes. But if your systems are not prepared, that interruption could corrupt databases, lose transactions in flight, crash applications, and wreak all manner of havoc. Their downtime glitch will become your costly business disaster unless you are prepared in advance to control it on your end (source).

Hosting.com's Service Level Agreement is posted publicly on their website (source). A quick read reveals that it does NOT promise 100% uptime. Datacenters fail, and that's a fact of life. When signing with any hosting or cloud provider, it is vital that you understand exactly who is responsible for what, and whether total downtime is measured according to the unavailability of their infrastructure or the amount of time it takes you to recover.

Recently, the Paris-based International Working Group on Cloud Computing Resiliency (IWGCR) found that costs for outages between 2007 and 2011 among the 13 providers they reviewed exceeded $70 million (source). No SLA from any provider is going to compensate for those kinds of losses. If the industry demanded this of them, no hosting provider would be able to stay in business. It will be far better for their customers to invest in reliability than to expect dollar-for-dollar restitution after a disaster.

Background: Hosting.com and its Customers

According to the company website, "Hosting.com is a next generation cloud hosting and recovery services company focused on ensuring your mission-critical applications are AlwaysOn™" (source). They are a leading provider of other enterprise hosting solutions and services as well, with datacenters in Dallas, Denver, Irvine, Louisville, Newark, and San Francisco. One source says they host over 65,000 websites (source). This includes financial services, healthcare, media, retail, software as a service (SaaS) providers, and content distribution networks (CDN) (source).

In contrast with other recent outages and other providers whose explanations were late or non-existent, Hosting.com CEO Art Zeile stepped up very quickly during this crisis and alerted his customers of the problem, its cause, and its effects. Though they won't be thrilled with news like this, customers need clear communication and honesty from their providers. That way they know what to tell their own customers and management, and their overworked internal IT teams will have a better chance of taming the chaos. I applaud Mr. Zeile and his actions. We need this level of leadership to benefit the cloud industry at large.

The Problem: Human Error, a Power Outage, and a Chain Reaction

Mr. Zeile explained that "An incorrect breaker operation sequence executed by the servicing vendor caused a shutdown of the UPS plant resulting in loss of critical power to one data center suite within the [Newark, Delaware] facility" (source). The power was back on within 11 minutes, but "customer web sites were offline for between one and five hours as their equipment and databases required more time to recover from the sudden loss of power."

I wasn't there but I can surmise what happened. When the power went out, an unspecified number of servers were shut off without proper shutdown procedures. Applications and databases were abruptly terminated. Other applications and databases that depended upon them suddenly lost transactions in flight. They crashed too, taking down other apps and databases in turn. And so on down the line in a classic cascading failure scenario.

Recovery of the customers' crashed apps and databases required hours. Each customer needed its own data and apps restored, and those that were still running probably had to be shut down and then re-started in proper sequence. Servers had to be checked for damage after their "crash" shut-downs. Apps and data that successfully cut over or failed-over to secondaries had to be cut over again, from the secondaries back to the primaries, and I'll bet there were further failures as that happened.

The Solution: Make Your Datacenters Expendable

Many of the apps and data on the system were undoubtedly protected by failover and backup recovery architecture, or by one of the Hosting.com business continuity solutions. Many of these certainly continued running as they successfully failed over to their secondaries. But equally clear is that apps and data for about 1,100 customers (1.7% of the total Hosting.com customer base) did not continue running. Either they were not equipped with adequate business continuity systems, or the failovers failed. One writer quotes Zeile as saying that although Hosting.com offers a backup option "few customers, at the affected location, had elected to purchase it" (source).

I am unaware of any hosting or cloud provider who publicly promises 100% uptime. So the customer must expect to have some amount of downtime, if only for maintenance. Logically, customers need to provide adequate business continuity systems to protect themselves.

Datacenters go offline all the time for any number of reasons. Thus, your business needs to be able to continue talking to customers, sending billing statements, shipping goods, and paying creditors despite untoward events like power outages, fires, human error, hardware failure, and so forth.

ZeroNines does not recommend or use a failover- or backup-based recovery paradigm. We take a different approach aimed at preventing downtime in the first place, rather than recovering from it afterward. In the case of the Hosting.com outage, Always Available™ architecture from ZeroNines offers two solution scenarios:
1)  Hosting.com already operates multiple geographically separated datacenters. Always Available architecture would allow processing to continue on any or all of the remaining five when any one of them goes down.
2)  Hosting.com customers could deploy their own Always Available array that would simultaneously replicate all transactions and data on other Hosting.com datacenters, or in other clouds or with other providers.

In either case, the end user experiences no downtime because the remaining nodes continue processing as usual. The offending datacenter simply drops out of the array until power is restored or until your staff can repair it. The other nodes of the Always Available array will update the damaged node once it is functioning again, and bring it to an identical logical state.

Visit the ZeroNines website to find out more about how our disaster-proof architecture protects businesses of any description from downtime.

Alan Gin – Founder & CEO, ZeroNines

The Cost of Cloud Outages and Planning for the Next One

At least one customer is publicly abandoning the Amazon EC2 cloud after two power-related outages within a month. Online dating site WhatsYourPrice.com is walking out, going back to a more traditional local hosting provider (source). And so the heat is still on in the East, with high temperatures still making life difficult and with the Amazon cloud beginning to lose customers.

What is the Real Cost of Cloud Outages?

This question is difficult to answer, as little hard data is available. Cloud providers and cloud customers keep the financial numbers closely guarded, whether we are talking about the customer's lost business and recovery costs, or Amazon's losses due to lost revenue, restitution paid to customers, customer attrition, and the increased difficulty of acquiring new customers.

Coincidentally, a report on cloud outages among top providers came out on June 22 2012. In it, the Paris-based International Working Group on Cloud Computing Resiliency (IWGCR) claims that costs for outages between 2007 and 2011 among the 13 providers they reviewed totaled $70 million. Their estimates were based on "hourly costs accepted in the industry" (source).

Downtime and availability rates are reported for Amazon Web Services (AWS), Microsoft, Research in Motion (RIM), and others. Total downtime was 568 hours, and availability was 99.917%, nowhere near the five nines (99.999%) that is becoming the de facto target for acceptable uptime.

Although it’s useful to put a line in the sand and publish studies on the cost of outages, such surveys are virtually impossible to do accurately. I don't think that even big analysts like Gartner are able to get a good view into the real costs. Unfortunately, this article does not make it clear whether the $70 million was the cost to providers, to customers, or both. Also, the sample size was pretty small and apparently there is no information about actual customer size. I know first-hand that many of our clients claim their downtime costs start at $6 million per hour and average $18-24 million per hour. Apparently only outages that made the news were included in the report, so this leaves a lot of actual downtime out of their equation, such as small glitches and maintenance downtime that journalists don't hear about. Because of all this I know the actual costs must be higher. Despite all the unavoidable barriers to accurate measurement, such studies are still valuable because they highlight the bottom-line impacts. They also demonstrate just how difficult it is to estimate the cost of downtime.

40% of Cloud Users are Not Prepared

"Light, medium, and heavy cloud users are running clouds where on average 40 percent of their cloud — data, applications, and infrastructure — is NOT backed up and exposed to outage meltdown" (source).

This was said a couple weeks ago by Cameron Peron, VP Marketing at Newvem, a cloud optimization consultancy that specializes in the Amazon cloud. He was referring to his company's clients and the June 15 outage.

If the average cloud customer is anything like these companies then it is no wonder that cloud outages are such a concern. Another writer referred to this kind of planning (or lack of planning) as "stupid IT mistakes" (source).

Who's at Fault?

So when a cloud customer experiences downtime and loses money, who is actually to blame? The cloud provider who failed to deliver 100% uptime, or the cloud customer who was unprepared for the unavoidable downtime?

According to Peron, "Amazon doesn’t make any promises to back up data... The real issue is that many users are under the impression that their data is backed up… but in fact it isn’t due to mismanaged infrastructure configuration." (source)

Cloud customers need to be prepared to use best practices for data protection and disaster prevention/recovery. They need to remember that a cloud is just a virtual datacenter. It is a building crammed full of servers, each of which is home to a number of virtual servers. And all of it is subject to the thousand natural (and unnatural) shocks that silicon is heir to.

Cloud Customers Need to Take Responsibility for Continuity

So here's some friendly advice to WhatsYourPrice.com and others like them: whatever hosting model you choose, get your disaster plan in place. An outage is out there, waiting for you in the form of a bad cooling fan, corrupt database, fire, flood, or human error whether it's in the cloud, your own virtual servers, or a local hosting provider.

Best practices and DR discipline should not be taken for granted simply because the datacenter is outsourced. Many companies that I meet with have gone to the cloud to cut costs, and many of those are reinvesting their savings into providing higher availability. They're looking ahead, trying to avoid disasters, outcompete based on performance, and support customer satisfaction.

Even before the advent of the cloud, new generations of low-cost compute models enabled disaster recovery standards that could prevent a lot of downtime. But they are often poorly executed or ignored altogether. And now, with it being so easy to outsource hosting to "the cloud", it is even easier for companies to shake off responsibility for business continuity, assuming or hoping the folks behind the curtain will take care of everything.

I say that the primary responsibility for outages is the customer's. If you're providing a high-demand service you need to be ready to deliver. It doesn't matter if you can legitimately blame your provider after a disaster; your customers will blame you.

Your cloud (or other hosting) provider no doubt promises a certain amount of uptime in their service level agreement. Let's imagine that allows one hour of downtime per year. If they have one minor problem it could cause downtime of just a few minutes. But if your systems are not prepared, that interruption could corrupt databases, lose transactions in flight, crash applications, and wreak all manner of havoc. Their downtime glitch will become your costly business disaster unless you are prepared in advance to control it on your end.

It's like a tire blowing out on a car; the manufacturer may be responsible to a degree for a wreck, but if your seatbelt was not fastened then all bets are off. Safety systems are there for a reason.

Make it so Any Datacenter is Expendable

ZeroNines offers a solution that enables the complete loss of any datacenter without causing service downtime. We believe that if WhatsYourPrice.com was using our Always Available™ architecture, their dating service would have continued to operate at full capacity for the duration of the outage, with no impact to the customer experience.

Visit the ZeroNines website to find out more about how our disaster-proof architecture protects businesses of any description from downtime. You may also be interested in our whitepaper "Cloud Computing Observations and the Value of CloudNines™".

Alan Gin – Founder & CEO, ZeroNines

Multi-Region Disasters and Expendable Databases℠

As I write this, the United States is suffering from a frightening heat wave and the lingering effects of storms that threatened everything east of the Rockies. Over two dozen lives have been lost and the heat might last for several more days. The transportation, emergency response, and utility infrastructures are badly strained; about a million customers are still without power. Major fires are burning in the west. It is distressing to think about this kind of multi-region disaster, but it is happening now and continues to unfold.

And as trivial as it feels to write this, these natural disasters led to another outage of the Amazon EC2 cloud on June 29. This one is very similar to the EC2 outage on June 15 which I blogged about last week (source). Friday's happened at the same Virginia datacenter and was also caused by the same kind of event.

The Problem: Another Power Outage and Generator Failure

Similarly to what happened a couple weeks ago, on Friday a storm-induced power outage at the utility company forced a switchover to Amazon's backup generator, and this generator failed (source). Netflix, Instagram, Pinterest, and others began to experience difficulties and outages. The problems lasted about an hour.

But rather than dissect this one outage, let's take a look at the larger issues surrounding downtime in the cloud.

Cloud Users Must Plan their Own Disaster Recovery

Wired.com had this to say about Friday's event:

In theory, big outages like this aren’t supposed to happen. Amazon is supposed to keep the data centers up and running – something it has become very good at… In reality, though, Amazon data centers have outages all the time. In fact, Amazon tells its customers to plan for this to happen, and to be ready to roll over to a new data center whenever there’s an outage. (source)

Long and short, cloud customers think that they have shed their responsibility for business continuity and handed it to the cloud provider. They're wrong, and Amazon has apparently admitted as much by telling its customers to make their own disaster recovery preparations.

"Stupid IT Mistakes"

Those are the lead words in the title of an article about the June 15 outage (source). In it, the author refers to statistics from cloud optimization firm Newvem that show that 40% of cloud users are not properly prepared for an outage. They don't have any kind of redundancy: they don't back up their data and they deploy to only one region. Frighteningly, this includes large companies as well as small.

Promoting a Failure of a Recovery Plan

Another problem is that Amazon has apparently told its customers to "be ready to roll over to a new data center" (source). This is tacit approval of failover-based disaster recovery systems. But as we saw with the June 15 outage, failovers fail all the time and cannot be relied upon to maintain continuity. In fact, they often contribute to outages.

As for regular backups, that's always a good idea. But a backup location can fail too, particularly if it is hit by the same disaster. And what happens with transactions that occurred after the last backup? Will a recovery based on these backups even succeed? And although backup may eventually get you running again, it can't prevent the costly downtime.

You Can't Prevent All Causes of Failure

I argue again and again that there is no way to prevent all the thousands of small and large errors that can conspire (singly or in combination) to knock out a datacenter, cloud node, or server. Generators, power supplies, bad cables, human error and any number of other small disasters can easily combine to make a big disaster. It's not practical to continually monitor all of these to circumvent every possible failure. IT is chaos theory personified; do all you can, but something's going to break.

Geographic Issues

As we are seeing this week, one disaster or group of disasters can span vast geographic areas. You need to plan your business continuity system so the same disaster can't affect everything. Companies that have located all their IT on the Eastern Seaboard should be sweating it this week, because it's conceivable that the heat wave and storms could cause simultaneous power outages from New York to Virginia to Florida. A primary site, failover site, and backup location could all go down at the same time.

The Real Solution: Geographically Separated Expendable Datacenters℠

Here at ZeroNines we've constructed our business continuity solution around a number of tenets, including:

• Service must continue despite the complete failure of any one datacenter.
• Geographical separation is key, to prevent one disaster from wiping out everything.
• Failover is not an option, because it is extremely unreliable.
• The solution must be easy and affordable so that the "40%" mentioned above actually use it.

Based on all this, we've developed Always Available™ architecture that enables multiple instances of the same applications, data, and transactions to run equally and simultaneously in multiple locations hundreds or thousands of miles apart. The system does not rely upon or include either failover or restoration from backup. Best of all, an entire datacenter could go offline at any moment and all transactions will simply continue to be processed at other datacenters with no interruption to service. It is affordable, OS agnostic, and operates with existing apps, databases, and infrastructures.

ZeroNines client ZenVault uses Always Available. They also host on Amazon EC2. During the outages, no extraordinary measures are needed. If the EC2 East node goes offline, the two other nodes (in Santa Clara and EC2 West) will continue running the service, and will restore the EC2 East node once it comes back online. ZenVault has had true 100% uptime since the day it launched in 2010.

Visit the ZeroNines website to find out more about how our disaster-proof architecture protects businesses of any description from downtime.

Alan Gin – Founder & CEO, ZeroNines

Outage at Amazon EC2 Virginia Illustrates the Value of the Expendable Datacenter℠

The Amazon EC2 cloud had a relatively minor outage a couple weeks ago, on June 14 2012. As it turns out, it happened in the same Virginia datacenter that spawned the April 2011 and August 2011 outages. I've been on the road but now that I look into it I see that it's actually a classic outage scenario and a classic example of cascading failure resulting from a failover. It also illustrates just why you need to plan for Expendable Datacenters℠.

Background: Amazon and Their Cloud Service

I blogged about Amazon's big outage last August (source), and described how large a role Amazon plays in the cloud world. I won't recap all that here, but I will say that among its clients are Netflix, Instagram, Reddit, Foursquare, Quora, Pinterest, parts of Salesforce.com, and ZeroNines client ZenVault.

The Problem: A Power Outage

According to the Amazon status page (source) "a cable fault in the high voltage Utility power distribution system" led to a power outage at the datacenter. Primary backup generators successfully kicked in but after about nine minutes "one of the generators overheated and powered off because of a defective cooling fan." Secondary backup power successfully kicked in but after about four minutes a circuit breaker opened because it had been "incorrectly configured." At this point, with no power at all, some customers went completely offline. Others that were using Amazon's multi-Availability Zone configurations stayed online bu seem to have suffered from impaired API calls, described below. Power was restored about half an hour after it was first lost.

Sites started recovering as soon as power was restored and most customers were back online about an hour after the whole episode began. But it is clear that many weren't really ready for business again because of the cascading effects of the initial interruption.

Subsequent Problems: Loss of In-Flight Transactions

The Amazon report says that when power came back on, some instances were "in an inconsistent state" and that they may have lost "in-flight writes." I interpret this to mean that when the system failed over to the backups, the backup servers were not synchronized with the primaries, resulting in lost transactions. This is typical of a failover disaster recovery system.

Another Subsequent Problem: Impaired API Calls

Additionally, during the power outage, API calls related to Amazon Elastic Block Store (EBS) volumes failed. Amazon sums up the effect beautifully: "The datastore that lost power did not fail cleanly, leaving the system unable to flip [failover] the datastore to its replicas in another Availability Zone." Here's a second failed failover within the same disaster.

My Compliments to Amazon EC2

In all seriousness, I truly commend Amazon for publicly posting such a detailed description of the disaster. It looks to me like they handled the disaster quickly and efficiently within the limitations of their system. Unfortunately that system is clearly not suited to the job at hand.

Amazon does a pretty good job at uptime. We (ZeroNines) use the Amazon EC2 cloud ourselves. But we hedge our bets by adding our own commercially available Always Available™ architecture to harden the whole thing against power outages and such. If this outage had affected our particular instances, we would not have experienced any downtime, inconsistency, failed transactions, or other ill effects.

One Solution for Three Problems

Always Available runs multiple instances of the same apps and data in multiple clouds, virtual servers, or other hosting environments. All are hot and all are active.

When the power failed in the first phase of this disaster, two or more identical Always Available nodes would have continued processing as normal. The initial power outage would not have caused service downtime because customers would have been served by the other nodes.

Secondly, those in-flight transactions would not have been lost because the other nodes would have continued processing them. With Always Available there is no failover and consequently no "dead air" when transactions can be lost.

Third, those failed EBS API calls would not have failed because again, they would have gone to the remaining fully functional nodes.

A big issue in this disaster was the "inconsistent state," or lack of synchronization between the primary and the failover servers. Within an Always Available architecture, there is no failover. Each server is continually updating and being updated by all other servers. Synchronization takes place constantly so when one node is taken out of the configuration the others simply proceed as normal, processing all transactions in order. When the failed server is brought back online, the others update it and bring it to the same logical state so it can begin processing again.

The Expendable Datacenter

Another thing I can't help but point out is the string of events that caused the outage in the first place. First a cable failure combines with a fan failure, and that combines with a circuit breaker failure. It's simple stuff that adds up into a disaster. Then software that can't synchronize. Given the complexities of the modern datacenter, how many possible combinations of points of failure are there? Thousands? Millions? I'll go on the record and say that there is no way to map all the possible failures, and no way to guard against them all individually. Its far better to accept the fact that servers, nodes or entire facilities will go down someday, and that you need to make the whole datacenter expendable without affecting performance. That's what ZeroNines does.

So if you're a cloud customer, take a look at ZeroNines. We can offer virtually 100% uptime whether you host in the cloud, on virtual servers, or in a typical hosted environment. And if you're a cloud provider, you can apply Always Available architecture to your service offering, avoiding disasters like this in the first place.

Check back in a few days and I'll write another post that looks at this outage from a business planning perspective.

Visit the ZeroNines website to find out more about how our disaster-proof architecture protects businesses of any description from downtime.

Alan Gin – Founder & CEO, ZeroNines