Welcome!

Cloud Expo Authors: Pat Romanski, Ignacio M. Llorente, Haim Koshchitzky, Liz McMillan, Gary Kaiser

Related Topics: Java, SOA & WOA, Adobe Flex, AJAX & REA, Apache

Java: Article

Why Averages Are Inadequate, and Percentiles Are Great

Averages are ineffective because they are too simplistic and one-dimensional

Anyone who ever monitored or analyzed an application uses or has used averages. They are simple to understand and calculate. We tend to ignore just how wrong the picture is that averages paint of the world. To emphasis the point let me give you a real-world example outside of the performance space that I read recently in a newspaper.

The article was explaining that the average salary in a certain region in Europe was 1900 Euro's (to be clear this would be quite good in that region!). However when looking closer they found out that the majority, namely 9 out of 10 people, only earned around 1000 Euros and one would earn 10.000 (I over simplified this of course, but you get the idea). If you do the math you will see that the average of this is indeed 1900, but we can all agree that this does not represent the "average" salary as we would use the word in day to day live. So now let's apply this thinking to application performance.

The Average Response Time
The average response time is by far the most commonly used metric in application performance management. We assume that this represents a "normal" transaction, however this would only be true if the response time is always the same (all transaction run at equal speed) or the response time distribution is roughly bell curved.

A Bell curve represents the "normal" distribution of response times in which the average and the median are the same. It rarely ever occurs in real applications

In a Bell Curve the average (mean) and median are the same. In other words observed performance would represent the majority (half or more than half) of the transactions.

In reality most applications have few very heavy outliers; a statistician would say that the curve has a long tail. A long tail does not imply many slow transactions, but few that are magnitudes slower than the norm.

This is a typical Response Time Distribution with few but heavy outliers - it has a long tail. The average here is dragged to the right by the long tail.

We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.

You can now argue that this is not a problem as long as the average doesn't look better than the median. I would disagree, but let's look at another real-world scenario experienced by many of our customers:

This is another typical Response Time Distribution. Here we have quite a few very fast transactions that drag the average to the left of the actual median

In this case a considerable percentage of transactions are very, very fast (10-20 percent), while the bulk of transactions are several times slower. The median would still tell us the true story, but the average all of a sudden looks a lot faster than most of our transactions actually are. This is very typical in search engines or when caches are involved - some transactions are very fast, but the bulk are normal. Another reason for this scenario are failed transactions, more specifically transactions that failed fast. Many real-world applications have a failure rate of 1-10 percent (due to user errors or validation errors). These failed transactions are often magnitudes faster than the real ones and consequently distorted an average.

Of course performance analysts are not stupid and regularly try to compensate with higher frequency charts (compensating by looking at smaller aggregates visually) and by taking in minimum and maximum observed response times. However we can often only do this if we know the application very well, those unfamiliar with the application might easily misinterpret the charts. Because of the depth and type of knowledge required for this, it's difficult to communicate your analysis to other people - think how many arguments between IT teams have been caused by this. And that's before we even begin to think about communicating with business stakeholders!

A better metric by far are percentiles, because they allow us to understand the distribution. But before we look at percentiles, let's take a look a key feature in every production monitoring solution: Automatic Baselining and Alerting.

Automatic Baselining and Alerting
In real-world environments, performance gets attention when it is poor and has a negative impact on the business and users. But how can we identify performance issues quickly to prevent negative effects? We cannot alert on every slow transaction, since there are always some. In addition, most operations teams have to maintain a large number of applications and are not familiar with all of them, so manually setting thresholds can be inaccurate, quite painful and time-consuming.

The industry has come up with a solution called Automatic Baselining. Baselining calculates out the "normal" performance and only alerts us when an application slows down or produces more errors than usual. Most approaches rely on averages and standard deviations.

Without going into statistical details, this approach again assumes that the response times are distributed over a bell curve:

The Standard Deviation represents 33% of all transactions with the mean as the middle. 2xStandard Deviation represents 66% and thus the majority, everything outside could be considered an outlier. However most real world scenarios are not bell curved...

Typically, transactions that are outside two times standard deviation are treated as slow and captured for analysis. An alert is raised if the average moves significantly. In a bell curve this would account for the slowest 16.5 percent (and you can of course adjust that); however; if the response time distribution does not represent a bell curve, it becomes inaccurate. We either end up with a lot of false positives (transactions that are a lot slower than the average but when looking at the curve lie within the norm) or we miss a lot of problems (false negatives). In addition if the curve is not a bell curve, then the average can differ a lot from the median; applying a standard deviation to such an average can lead to quite a different result than you would expect. To work around this problem these algorithms have many tunable variables and a lot of "hacks" for specific use cases.

Why I Love Percentiles
A percentile tells me which part of the curve I am looking at and how many transactions are represented by that metric. To visualize this look at the following chart:

This chart shows the 50th and 90th percentile along with the average of the same transaction. It shows that the average is influenced far mor heavily by the 90th, thus by outliers and not by the bulk of the transactions

The green line represents the average. As you can see it is very volatile. The other two lines represent the 50th and 90th percentile. As we can see the 50th percentile (or median) is rather stable but has a couple of jumps. These jumps represent real performance degradation for the majority (50%) of the transactions. The 90th percentile (this is the start of the "tail") is a lot more volatile, which means that the outliers slowness depends on data or user behavior. What's important here is that the average is heavily influenced (dragged) by the 90th percentile, the tail, rather than the bulk of the transactions.

If the 50th percentile (median) of a response time is 500ms that means that 50% of my transactions are either as fast or faster than 500ms. If the 90th percentile of the same transaction is at 1000ms it means that 90% are as fast or faster and only 10% are slower. The average in this case could either be lower than 500ms (on a heavy front curve), a lot higher (long tail) or somewhere in between. A percentile gives me a much better sense of my real world performance, because it shows me a slice of my response time curve.

For exactly that reason percentiles are perfect for automatic baselining. If the 50th percentile moves from 500ms to 600ms I know that 50% of my transactions suffered a 20% performance degradation. You need to react to that.

In many cases we see that the 75th or 90th percentile does not change at all in such a scenario. This means the slow transactions didn't get any slower, only the normal ones did. Depending on how long your tail is the average might not have moved at all in such a scenario.!

In other cases we see the 98th percentile degrading from 1s to 1.5 seconds while the 95th is stable at 900ms. This means that your application as a whole is stable, but a few outliers got worse, nothing to worry about immediately. Percentile-based alerts do not suffer from false positives, are a lot less volatile and don't miss any important performance degradations! Consequently a baselining approach that uses percentiles does not require a lot of tuning variables to work effectively.

The screenshot below shows the Median (50th Percentile) for a particular transaction jumping from about 50ms to about 500ms and triggering an alert as it is significantly above the calculated baseline (green line). The chart labeled "Slow Response Time" on the other hand shows the 90th percentile for the same transaction. These "outliers" also show an increase in response time but not significant enough to trigger an alert.

Here we see an automatic baselining dashboard with a violation at the 50th percentile. The violation is quite clear, at the same time the 90th percentile (right upper chart) does not violate. Because the outliers are so much slower than the bulk of the transaction an average would have been influenced by them and would not have have reacted quite as dramatically as the 50th percentile. We might have missed this clear violation!

How Can We Use Percentiles for Tuning?
Percentiles are also great for tuning, and giving your optimizations a particular goal. Let's say that something within my application is too slow in general and I need to make it faster. In this case I want to focus on bringing down the 90th percentile. This would ensure sure that the overall response time of the application goes down. In other cases I have unacceptably long outliers I want to focus on bringing down response time for transactions beyond the 98th or 99th percentile (only outliers). We see a lot of applications that have perfectly acceptable performance for the 90th percentile, with the 98th percentile being magnitudes worse.

In throughput oriented applications on the other hand I would want to make the majority of my transactions very fast, while accepting that an optimization makes a few outliers slower. I might therefore make sure that the 75th percentile goes down while trying to keep the 90th percentile stable or not getting a lot worse.

I could not make the same kind of observations with averages, minimum and maximum, but with percentiles they are very easy indeed.

Conclusion
Averages are ineffective because they are too simplistic and one-dimensional. Percentiles are a really great and easy way of understanding the real performance characteristics of your application. They also provide a great basis for automatic baselining, behavioral learning and optimizing your application with a proper focus. In short, percentiles are great!

More Stories By Michael Kopp

Michael Kopp has over 12 years of experience as an architect and developer in the Enterprise Java space. Before coming to CompuwareAPM dynaTrace he was the Chief Architect at GoldenSource, a major player in the EDM space. In 2009 he joined dynaTrace as a technology strategist in the center of excellence. He specializes application performance management in large scale production environments with special focus on virtualized and cloud environments. His current focus is how to effectively leverage BigData Solutions and how these technologies impact and change the application landscape.

Comments (1) View Comments

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


Most Recent Comments
rtalexander 11/21/12 12:58:00 AM EST

Hey, could you post a reference or two that covers the theory and/or practicalities of the approach you describe?

Thanks!

Cloud Expo Breaking News
Cloud Computing is evolving into a Big Three of Amazon Web Services, Google Cloud, and Microsoft Azure. Cloud 360: Multi-Cloud Bootcamp, being held Nov 4–5, 2014, in conjunction with 15th Cloud Expo in Santa Clara, CA, delivers a real-world demonstration of how to deploy and configure a scalable and available web application on all three platforms. The Cloud 360 Bootcamp, led by Janakiram MSV, an analyst with Gigaom Research, is the first bootcamp that introduces the core concepts of Infrastructure as a Service (IaaS) based on the workings of the Big Three platforms – Amazon EC2, Google Compute Engine, and Azure VMs. Bootcamp attendees will get to see the big picture and also receive the knowledge needed to make the best cloud decisions for their business applications and entire enterprise IT organization.
Scott Jenson leads a project called The Physical Web within the Chrome team at Google. Project members are working to take the scalability and openness of the web and use it to talk to the exponentially exploding range of smart devices. Nearly every company today working on the IoT comes up with the same basic solution: use my server and you'll be fine. But if we really believe there will be trillions of these devices, that just can't scale. We need a system that is open a scalable and by using the URL as a basic building block, we open this up and get the same resilience that the web enjoys.
The Internet of Things is a natural complement to the cloud and related technologies such as Big Data, analytics, and mobility. In his session at Internet of @ThingsExpo, Joe Weinman will lay out four generic strategies – digital disciplines – to exploit emerging digital technologies for strategic advantage. Joe Weinman has held executive leadership positions at Bell Labs, AT&T, Hewlett-Packard, and Telx, in areas such as corporate strategy, business development, product management, operations, and R&D.
SYS-CON Events announced today that DevOps.com has been named “Media Sponsor” of SYS-CON's “DevOps Summit at Cloud Expo,” which will take place on June 10–12, 2014, at the Javits Center in New York City, New York. DevOps.com is where the world meets DevOps. It is the largest collection of original content relating to DevOps on the web today Featuring up-to-the-minute news, feature stories, blogs, bylined articles and more, DevOps.com is where the thought leaders of the DevOps movement make their ideas known.
There are 182 billion emails sent every day, generating a lot of data about how recipients and ISPs respond. Many marketers take a more-is-better approach to stats, preferring to have the ability to slice and dice their email lists based numerous arbitrary stats. However, fundamentally what really matters is whether or not sending an email to a particular recipient will generate value. Data Scientists can design high-level insights such as engagement prediction models and content clusters that allow marketers to cut through the noise and design their campaigns around strong, predictive signals, rather than arbitrary statistics. SendGrid sends up to half a billion emails a day for customers such as Pinterest and GitHub. All this email adds up to more text than produced in the entire twitterverse. We track events like clicks, opens and deliveries to help improve deliverability for our customers – adding up to over 50 billion useful events every month. While SendGrid data covers only abo...
SYS-CON Events announced today that the Web Host Industry Review has been named “Media Sponsor” of SYS-CON's 15th International Cloud Expo®, which will take place on November 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA. Since 2000, The Web Host Industry Review has made a name for itself as the foremost authority of the Web hosting industry providing reliable, insightful and comprehensive news, reviews and resources to the hosting community. TheWHIR Blogs provides a community of expert industry perspectives. The Web Host Industry Review Magazine also offers a business-minded, issue-driven perspective of interest to executives and decision-makers. WHIR TV offers on demand web hosting video interviews and web hosting video features of the key persons and events of the web hosting industry. WHIR Events brings together like-minded hosting industry professionals and decision-makers in local communities. TheWHIR is an iNET Interactive property.
SYS-CON Events announced today that O'Reilly Media has been named “Media Sponsor” of SYS-CON's 15th International Cloud Expo®, which will take place on November 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA. O'Reilly Media spreads the knowledge of innovators through its books, online services, magazines, and conferences. Since 1978, O'Reilly Media has been a chronicler and catalyst of cutting-edge development, homing in on the technology trends that really matter and spurring their adoption by amplifying "faint signals" from the alpha geeks who are creating the future. An active participant in the technology community, the company has a long history of advocacy, meme-making, and evangelism.
SYS-CON Events announced today that Verizon has been named “Gold Sponsor” of SYS-CON's 15th International Cloud Expo®, which will take place on November 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA. Verizon Enterprise Solutions creates global connections that generate growth, drive business innovation and move society forward. With industry-specific solutions and a full range of global wholesale offerings provided over the company's secure mobility, cloud, strategic networking and advanced communications platforms, Verizon Enterprise Solutions helps open new opportunities around the world for innovation, investment and business transformation. Visit verizonenterprise.com to learn more.
SYS-CON Events announced today that TMCnet has been named “Media Sponsor” of SYS-CON's 15th International Cloud Expo®, which will take place on November 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA. Technology Marketing Corporation (TMC) is the world's leading business to business and integrated marketing media company, servicing niche markets within the communications and technology industries.
"In my session I spoke about enterprise cloud analytics and how we can leverage analytics as a service," explained Ajay Budhraja, CTO at the Department of Justice, in this SYS-CON.tv interview at the 14th International Cloud Expo®, held June 10-12, 2014, at the Javits Center in New York City. Cloud Expo® 2014 Silicon Valley, November 4–6, at the Santa Clara Convention Center in Santa Clara, CA, will feature technical sessions from a rock star conference faculty and the leading Cloud industry players in the world.
“We are starting to see people move beyond the commodity cloud and enterprises need to start focusing on additional value added services in order to really drive their adoption," explained Jason Mondanaro, Director of Product Management at MetraTech, in this SYS-CON.tv interview at the 14th International Cloud Expo®, held June 10-12, 2014, at the Javits Center in New York City. Cloud Expo® 2014 Silicon Valley, November 4–6, at the Santa Clara Convention Center in Santa Clara, CA, will feature technical sessions from a rock star conference faculty and the leading Cloud industry players in the world.
"We are automated capacity control software, which basically looks at all the supply and demand and running a virtual cloud environment and does a deep analysis of that and says where should things go," explained Andrew Hillier, Co-founder & CTO of CiRBA, in this SYS-CON.tv interview at the 14th International Cloud Expo®, held June 10-12, 2014, at the Javits Center in New York City. Cloud Expo® 2014 Silicon Valley, November 4–6, at the Santa Clara Convention Center in Santa Clara, CA, will feature technical sessions from a rock star conference faculty and the leading Cloud industry players in the world.
Almost everyone sees the potential of Internet of Things but how can businesses truly unlock that potential. The key will be in the ability to discover business insight in the midst of an ocean of Big Data generated from billions of embedded devices via Systems of Discover. Businesses will also need to ensure that they can sustain that insight by leveraging the cloud for global reach, scale and elasticity. In his session at Internet of @ThingsExpo, Mac Devine, Distinguished Engineer at IBM, will discuss bringing these three elements together via Systems of Discover.
The Internet of Things promises to transform businesses (and lives), but navigating the business and technical path to success can be difficult to understand. In his session at 15th Internet of @ThingsExpo, Chad Jones, Vice President, Product Strategy of LogMeIn's Xively IoT Platform, will show you how to approach creating broadly successful connected customer solutions using real world business transformation studies including New England BioLabs and more.
All too many discussions about DevOps conclude that the solution is an all-purpose player: developer and operations guru, complete with pager for round-the-clock duty. For most organizations that is not the way forward. In his session at DevOps Summit, Bernard Golden, Vice President of Strategy at ActiveState, will discuss how to achieve the agility and speed of end-to-end automation without requiring an organization stocked with Supermen and Superwomen.