Welcome!

Related Topics: Java IoT, Microservices Expo, Adobe Flex, Machine Learning , Apache

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We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.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

Latest Stories
Cloud resources, although available in abundance, are inherently volatile. For transactional computing, like ERP and most enterprise software, this is a challenge as transactional integrity and data fidelity is paramount – making it a challenge to create cloud native applications while relying on RDBMS. In his session at 21st Cloud Expo, Claus Jepsen, Chief Architect and Head of Innovation Labs at Unit4, will explore that in order to create distributed and scalable solutions ensuring high availa...
Cloud adoption is often driven by a desire to increase efficiency, boost agility and save money. All too often, however, the reality involves unpredictable cost spikes and lack of oversight due to resource limitations. In his session at 20th Cloud Expo, Joe Kinsella, CTO and Founder of CloudHealth Technologies, tackled the question: “How do you build a fully optimized cloud?” He will examine: Why TCO is critical to achieving cloud success – and why attendees should be thinking holistically ab...
In IT, we sometimes coin terms for things before we know exactly what they are and how they’ll be used. The resulting terms may capture a common set of aspirations and goals – as “cloud” did broadly for on-demand, self-service, and flexible computing. But such a term can also lump together diverse and even competing practices, technologies, and priorities to the point where important distinctions are glossed over and lost.
Internet-of-Things discussions can end up either going down the consumer gadget rabbit hole or focused on the sort of data logging that industrial manufacturers have been doing forever. However, in fact, companies today are already using IoT data both to optimize their operational technology and to improve the experience of customer interactions in novel ways. In his session at @ThingsExpo, Gordon Haff, Red Hat Technology Evangelist, shared examples from a wide range of industries – including en...
Detecting internal user threats in the Big Data eco-system is challenging and cumbersome. Many organizations monitor internal usage of the Big Data eco-system using a set of alerts. This is not a scalable process given the increase in the number of alerts with the accelerating growth in data volume and user base. Organizations are increasingly leveraging machine learning to monitor only those data elements that are sensitive and critical, autonomously establish monitoring policies, and to detect...
"We're a cybersecurity firm that specializes in engineering security solutions both at the software and hardware level. Security cannot be an after-the-fact afterthought, which is what it's become," stated Richard Blech, Chief Executive Officer at Secure Channels, in this SYS-CON.tv interview at @ThingsExpo, held November 1-3, 2016, at the Santa Clara Convention Center in Santa Clara, CA.
Most companies are adopting or evaluating container technology - Docker in particular - to speed up application deployment, drive down cost, ease management and make application delivery more flexible overall. As with most new architectures, this dream takes a lot of work to become a reality. Even when you do get your application componentized enough and packaged properly, there are still challenges for DevOps teams to making the shift to continuous delivery and achieving that reduction in cost ...
Enterprise architects are increasingly adopting multi-cloud strategies as they seek to utilize existing data center assets, leverage the advantages of cloud computing and avoid cloud vendor lock-in. This requires a globally aware traffic management strategy that can monitor infrastructure health across data centers and end-user experience globally, while responding to control changes and system specification at the speed of today’s DevOps teams. In his session at 20th Cloud Expo, Josh Gray, Chie...
To get the most out of their data, successful companies are not focusing on queries and data lakes, they are actively integrating analytics into their operations with a data-first application development approach. Real-time adjustments to improve revenues, reduce costs, or mitigate risk rely on applications that minimize latency on a variety of data sources. Jack Norris reviews best practices to show how companies develop, deploy, and dynamically update these applications and how this data-first...
Intelligent Automation is now one of the key business imperatives for CIOs and CISOs impacting all areas of business today. In his session at 21st Cloud Expo, Brian Boeggeman, VP Alliances & Partnerships at Ayehu, will talk about how business value is created and delivered through intelligent automation to today’s enterprises. The open ecosystem platform approach toward Intelligent Automation that Ayehu delivers to the market is core to enabling the creation of the self-driving enterprise.
"At the keynote this morning we spoke about the value proposition of Nutanix, of having a DevOps culture and a mindset, and the business outcomes of achieving agility and scale, which everybody here is trying to accomplish," noted Mark Lavi, DevOps Solution Architect at Nutanix, in this SYS-CON.tv interview at @DevOpsSummit at 20th Cloud Expo, held June 6-8, 2017, at the Javits Center in New York City, NY.
"We're here to tell the world about our cloud-scale infrastructure that we have at Juniper combined with the world-class security that we put into the cloud," explained Lisa Guess, VP of Systems Engineering at Juniper Networks, in this SYS-CON.tv interview at 20th Cloud Expo, held June 6-8, 2017, at the Javits Center in New York City, NY.
Historically, some banking activities such as trading have been relying heavily on analytics and cutting edge algorithmic tools. The coming of age of powerful data analytics solutions combined with the development of intelligent algorithms have created new opportunities for financial institutions. In his session at 20th Cloud Expo, Sebastien Meunier, Head of Digital for North America at Chappuis Halder & Co., discussed how these tools can be leveraged to develop a lasting competitive advantage ...
In his session at 20th Cloud Expo, Mike Johnston, an infrastructure engineer at Supergiant.io, discussed how to use Kubernetes to set up a SaaS infrastructure for your business. Mike Johnston is an infrastructure engineer at Supergiant.io with over 12 years of experience designing, deploying, and maintaining server and workstation infrastructure at all scales. He has experience with brick and mortar data centers as well as cloud providers like Digital Ocean, Amazon Web Services, and Rackspace. H...
The question before companies today is not whether to become intelligent, it’s a question of how and how fast. The key is to adopt and deploy an intelligent application strategy while simultaneously preparing to scale that intelligence. In her session at 21st Cloud Expo, Sangeeta Chakraborty, Chief Customer Officer at Ayasdi, will provide a tactical framework to become a truly intelligent enterprise, including how to identify the right applications for AI, how to build a Center of Excellence to ...