Welcome!

XML Authors: Trevor Parsons, Lacey Thoms, Elizabeth White, Carmen Gonzalez, Pat Romanski

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We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.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

@ThingsExpo Stories
Software AG helps organizations transform into Digital Enterprises, so they can differentiate from competitors and better engage customers, partners and employees. Using the Software AG Suite, companies can close the gap between business and IT to create digital systems of differentiation that drive front-line agility. We offer four on-ramps to the Digital Enterprise: alignment through collaborative process analysis; transformation through portfolio management; agility through process automation and integration; and visibility through intelligent business operations and big data.
There will be 50 billion Internet connected devices by 2020. Today, every manufacturer has a propriety protocol and an app. How do we securely integrate these "things" into our lives and businesses in a way that we can easily control and manage? Even better, how do we integrate these "things" so that they control and manage each other so our lives become more convenient or our businesses become more profitable and/or safe? We have heard that the best interface is no interface. In his session at Internet of @ThingsExpo, Chris Matthieu, Co-Founder & CTO at Octoblu, Inc., will discuss how these devices generate enough data to learn our behaviors and simplify/improve our lives. What if we could connect everything to everything? I'm not only talking about connecting things to things but also systems, cloud services, and people. Add in a little machine learning and artificial intelligence and now we have something interesting...
Last week, while in San Francisco, I used the Uber app and service four times. All four experiences were great, although one of the drivers stopped for 30 seconds and then left as I was walking up to the car. He must have realized I was a blogger. None the less, the next car was just a minute away and I suffered no pain. In this article, my colleague, Ved Sen, Global Head, Advisory Services Social, Mobile and Sensors at Cognizant shares his experiences and insights.
We are reaching the end of the beginning with WebRTC and real systems using this technology have begun to appear. One challenge that faces every WebRTC deployment (in some form or another) is identity management. For example, if you have an existing service – possibly built on a variety of different PaaS/SaaS offerings – and you want to add real-time communications you are faced with a challenge relating to user management, authentication, authorization, and validation. Service providers will want to use their existing identities, but these will have credentials already that are (hopefully) irreversibly encoded. In his session at Internet of @ThingsExpo, Peter Dunkley, Technical Director at Acision, will look at how this identity problem can be solved and discuss ways to use existing web identities for real-time communication.
Can call centers hang up the phones for good? Intuitive Solutions did. WebRTC enabled this contact center provider to eliminate antiquated telephony and desktop phone infrastructure with a pure web-based solution, allowing them to expand beyond brick-and-mortar confines to a home-based agent model. It also ensured scalability and better service for customers, including MUY! Companies, one of the country's largest franchise restaurant companies with 232 Pizza Hut locations. This is one example of WebRTC adoption today, but the potential is limitless when powered by IoT. Attendees will learn real-world benefits of WebRTC and explore future possibilities, as WebRTC and IoT intersect to improve customer service.
From telemedicine to smart cars, digital homes and industrial monitoring, the explosive growth of IoT has created exciting new business opportunities for real time calls and messaging. In his session at Internet of @ThingsExpo, Ivelin Ivanov, CEO and Co-Founder of Telestax, will share some of the new revenue sources that IoT created for Restcomm – the open source telephony platform from Telestax. Ivelin Ivanov is a technology entrepreneur who founded Mobicents, an Open Source VoIP Platform, to help create, deploy, and manage applications integrating voice, video and data. He is the co-founder of TeleStax, an Open Source Cloud Communications company that helps the shift from legacy IN/SS7 telco networks to IP-based cloud comms. An early investor in multiple start-ups, he still finds time to code for his companies and contribute to open source projects.
The Internet of Things (IoT) promises to create new business models as significant as those that were inspired by the Internet and the smartphone 20 and 10 years ago. What business, social and practical implications will this phenomenon bring? That's the subject of "Monetizing the Internet of Things: Perspectives from the Front Lines," an e-book released today and available free of charge from Aria Systems, the leading innovator in recurring revenue management.
The Internet of Things will put IT to its ultimate test by creating infinite new opportunities to digitize products and services, generate and analyze new data to improve customer satisfaction, and discover new ways to gain a competitive advantage across nearly every industry. In order to help corporate business units to capitalize on the rapidly evolving IoT opportunities, IT must stand up to a new set of challenges.
There’s Big Data, then there’s really Big Data from the Internet of Things. IoT is evolving to include many data possibilities like new types of event, log and network data. The volumes are enormous, generating tens of billions of logs per day, which raise data challenges. Early IoT deployments are relying heavily on both the cloud and managed service providers to navigate these challenges. In her session at 6th Big Data Expo®, Hannah Smalltree, Director at Treasure Data, to discuss how IoT, Big Data and deployments are processing massive data volumes from wearables, utilities and other machines.
All major researchers estimate there will be tens of billions devices – computers, smartphones, tablets, and sensors – connected to the Internet by 2020. This number will continue to grow at a rapid pace for the next several decades. With major technology companies and startups seriously embracing IoT strategies, now is the perfect time to attend @ThingsExpo in Silicon Valley. Learn what is going on, contribute to the discussions, and ensure that your enterprise is as "IoT-Ready" as it can be!
P2P RTC will impact the landscape of communications, shifting from traditional telephony style communications models to OTT (Over-The-Top) cloud assisted & PaaS (Platform as a Service) communication services. The P2P shift will impact many areas of our lives, from mobile communication, human interactive web services, RTC and telephony infrastructure, user federation, security and privacy implications, business costs, and scalability. In his session at Internet of @ThingsExpo, Erik Lagerway, Co-founder of Hookflash, will walk through the shifting landscape of traditional telephone and voice services to the modern P2P RTC era of OTT cloud assisted services.
While great strides have been made relative to the video aspects of remote collaboration, audio technology has basically stagnated. Typically all audio is mixed to a single monaural stream and emanates from a single point, such as a speakerphone or a speaker associated with a video monitor. This leads to confusion and lack of understanding among participants especially regarding who is actually speaking. Spatial teleconferencing introduces the concept of acoustic spatial separation between conference participants in three dimensional space. This has been shown to significantly improve comprehension and conference efficiency.
The Internet of Things is tied together with a thin strand that is known as time. Coincidentally, at the core of nearly all data analytics is a timestamp. When working with time series data there are a few core principles that everyone should consider, especially across datasets where time is the common boundary. In his session at Internet of @ThingsExpo, Jim Scott, Director of Enterprise Strategy & Architecture at MapR Technologies, will discuss single-value, geo-spatial, and log time series data. By focusing on enterprise applications and the data center, he will use OpenTSDB as an example to explain some of these concepts including when to use different storage models.
SYS-CON Events announced today that Gridstore™, the leader in software-defined storage (SDS) purpose-built for Windows Servers and Hyper-V, will exhibit at 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. Gridstore™ is the leader in software-defined storage purpose built for virtualization that is designed to accelerate applications in virtualized environments. Using its patented Server-Side Virtual Controller™ Technology (SVCT) to eliminate the I/O blender effect and accelerate applications Gridstore delivers vmOptimized™ Storage that self-optimizes to each application or VM across both virtual and physical environments. Leveraging a grid architecture, Gridstore delivers the first end-to-end storage QoS to ensure the most important App or VM performance is never compromised. The storage grid, that uses Gridstore’s performance optimized nodes or capacity optimized nodes, starts with as few a...
The Transparent Cloud-computing Consortium (abbreviation: T-Cloud Consortium) will conduct research activities into changes in the computing model as a result of collaboration between "device" and "cloud" and the creation of new value and markets through organic data processing High speed and high quality networks, and dramatic improvements in computer processing capabilities, have greatly changed the nature of applications and made the storing and processing of data on the network commonplace. These technological reforms have not only changed computers and smartphones, but are also changing the data processing model for all information devices. In particular, in the area known as M2M (Machine-To-Machine), there are great expectations that information with a new type of value can be produced using a variety of devices and sensors saving/sharing data via the network and through large-scale cloud-type data processing. This consortium believes that attaching a huge number of devic...
Innodisk is a service-driven provider of industrial embedded flash and DRAM storage products and technologies, with a focus on the enterprise, industrial, aerospace, and defense industries. Innodisk is dedicated to serving their customers and business partners. Quality is vitally important when it comes to industrial embedded flash and DRAM storage products. That’s why Innodisk manufactures all of their products in their own purpose-built memory production facility. In fact, they designed and built their production center to maximize manufacturing efficiency and guarantee the highest quality of our products.
Can call centers hang up the phones for good? Intuitive Solutions did. WebRTC enabled this contact center provider to eliminate antiquated telephony and desktop phone infrastructure with a pure web-based solution, allowing them to expand beyond brick-and-mortar confines to a home-based agent model. Download Slide Deck: ▸ Here
All major researchers estimate there will be tens of billions devices - computers, smartphones, tablets, and sensors - connected to the Internet by 2020. This number will continue to grow at a rapid pace for the next several decades. Over the summer Gartner released its much anticipated annual Hype Cycle report and the big news is that Internet of Things has now replaced Big Data as the most hyped technology. Indeed, we're hearing more and more about this fascinating new technological paradigm. Every other IT news item seems to be about IoT and its implications on the future of digital business.
BSQUARE is a global leader of embedded software solutions. We enable smart connected systems at the device level and beyond that millions use every day and provide actionable data solutions for the growing Internet of Things (IoT) market. We empower our world-class customers with our products, services and solutions to achieve innovation and success. For more information, visit www.bsquare.com.
With the iCloud scandal seemingly in its past, Apple announced new iPhones, updates to iPad and MacBook as well as news on OSX Yosemite. Although consumers will have to wait to get their hands on some of that new stuff, what they can get is the latest release of iOS 8 that Apple made available for most in-market iPhones and iPads. Originally announced at WWDC (Apple’s annual developers conference) in June, iOS 8 seems to spearhead Apple’s newfound focus upon greater integration of their products into everyday tasks, cross-platform mobility and self-monitoring. Before you update your device, here is a look at some of the new features and things you may want to consider from a mobile security perspective.