Welcome!

Machine Learning Authors: William Schmarzo, Kevin Jackson, Stackify Blog, Elizabeth White, Pat Romanski

Related Topics: Java IoT, Microservices Expo, Microsoft Cloud, Machine Learning , Agile Computing, @DXWorldExpo

Java IoT: Article

Eating Our Own Dog Food – 2x Faster Hadoop MapReduce Jobs

How to analyze and optimize Hadoop jobs beyond just tweaking MapReduce options

For a while now I have been writing about how to analyze and optimize Hadoop jobs beyond just tweaking MapReduce options. The other day I took a look at some of our Outage Analyzer Hadoop jobs and put words into action.

A simple analysis of the Outage Analyzer jobs with Compuware APM 5.5 identified three hotspots and two potential Hadoop problems in one of our biggest jobs. It took the responsible developer a couple of hours to fix it and the result is a 2x improvement overall and a 6x improvement on the Reduce part of the job. Let's see how we achieved that.

About Outage Analyzer
Outage Analyzer is a free service provided by Compuware that displays in real-time any availability problems with the most popular third-party content providers on the Internet. It's available at http://www.outageanalyzer.com. It uses real time analytical process technologies to do anomaly detection and event correlation and classification. It stores billions of measures taken from Compuware's global testing network every day in Hadoop and runs different MapReduce jobs to analyze the data. I examined the performance of these MapReduce jobs.

Identifying Worthwhile Jobs to Analyze
The first thing I did was look for a worthwhile job to analyze. To do this, I looked at cluster utilization broken down by user and job.

This chart visualizes the cluster CPU usage by user giving a good indication about which user executes the most expensive jobs

What I found was that John was the biggest user of our cluster. So I looked at the jobs John was running.

These are all the jobs that John was running over the last several days. It's always the same one, consuming about the same amount of resources

The largest job by far was an analytics simulation of a full day of measurement data. This job is run often to test and tune changes to the analytics algorithms. Except for one of the runs, all of them lasted about 6.5 hours in real time and consumed nearly 100% CPU of the cluster during that time. This looked like a worthy candidate for further analysis.

Identifying Which Job Phase to Focus On
My next step was to look at a breakdown of the job from two angles: consumption and real time. From a real-time perspective, map and reduce took about the same amount of time - roughly 3 hours each. This could also be nicely seen in the resource usage of the job.

This dashboard shows the overall cluster utilization during the time the job ran. 100% of the cluster CPU is used during most of the time

The significant drop in Load Average and the smaller drop in the other charts mark the end of the mapping phase and the start of pure reducing. What is immediately obvious is that the reducing phase, while lasting about the same time, does not consume as many resources as the map phase. The Load Average is significantly lower and the CPU utilization drops in several steps before the job is finished.

On the one hand that is because we have a priority scheduler and reducing does not use all slots, but more important, reducing cannot be parallelized as much as mapping. Every optimization here counts twofold, because you cannot scale things away. While the mapping phase is clearly consuming more resources, the reducing phase is a bottleneck and might therefore benefit even more from optimization.

The breakdown of job phase times shows that the mapping phase consumes twice as much time as reducing, even though we know that the job real time of the two phases is about the same - 3 hours each

As we can see the Map Time (time we spend in the mapping function, excluding merging, spilling and combining) is twice as high as the reduce time. The reduce time here represents the time that tasks were actually spending in the reduce function, excluding shuffle and merge time (which is represented separately). As such those two times represent those portions of the job that are directly impacted by the Map and Reduce code, which is usually custom code - and therefore tuneable by our developers.

Analyzing Map and Reduce Performance
As a next step I used Compuware APM to get a high-level performance breakdown of the job's respective 3 hour mapping and reducing phases. A single click gave me this pretty clear picture of the mapping phase:

This is a hot spot analysis of our 3 hour mapping phase which ran across 10 servers in our hadoop cluster

The hot spot dashboard for the mapping phase shows that we spent the majority of the time (about 70%) in our own code and that it's about 50% CPU time. This indicates a lot of potential for improvement. Next, I looked at the reducing phase.

This hot spot shows that we spend nearly all of our reducing time in all reduce tasks in our own code.

This shows that 99% of the reducing time is spent on our own code and not in any Hadoop framework. Since the reduce phase was clearly the winner in terms of potential, I looked at the details of that hot spot - and immediately found three hot spots that were responsible for the majority of the reduce time.

Three Simple Code Issues Consume 70% of the Reduce Time
This is what the method hot spots dashboard for the reduce phase showed.

These are the method hot spots for the reduce phase, showing that nearly everything is down to only three line items

The top three items in the method hot spot told me everything I needed know. As it turned out nearly all other items listed were sub-hotspots of the top most method:

  1. SimpleDateFormat initialization:
    The five items marked in red are all due to the creation of a SimpleDateFormat object. As most of us find out very painfully during our early career as a Java developer, the SimpleDateFormat is not thread safe and cannot be used as a static variable easily. This is why the developer chose the easiest route and created a new one for every call, leading to about 1.5 billion creations of this object. The initialization of the Formatter is actually very expensive and involves resource lookups, locale lookups and time calculations (seen in the separate line items listed here). This item alone consumed about 40% of our reduce time.
    Solution: We chose to use the well-known Joda framework (the code replacement was easy) and made the Formatter a static final variable; totally removing this big hot spot from the equation.
  2. Regular Expression Matching (line two in the picture)
    In order to split the CSV input we were using java.lang.String.split. It is often forgotten that this method uses regular expressions underneath. RegEx is rather CPU intensive and overkill for such a simple job. This was consuming another 15-20% of the allotted CPU time.
    Solution: We changed this to a simple string tokenizer.
  3. Exception throwing (line three in the picture)
    This example was especially interesting. During the reading of input data we are parsing numeric values, and if the field is not a correct number java.lang.Long.parseLong will throw a NumberFormatException. Our code would catch it, mark the field as invalid and ignore the exception. The fact is that nearly every input record in that feed has an invalid field or an empty field that should contain a number. Throwing this exception billions of times consumed another 10% of our CPU time.
    Solution: We changed the code in a way to avoid the exception altogether.

There we have it - three simple hot spots were consuming about 70% of our reduce CPU time. During analysis of the mapping portion I found the same hot spots again, where they contributed about 20-30% to the CPU time.

I sent this analysis to the developer and we decided to eat our own dog food, fix it and rerun the job to analyze the changes.

Job Done in Half the Time - Sixfold Improvement in Reduce Time
The result of the modified code exceeded our expectations by quite a bit. The immediate changes saw the job time reduced by 50%. Instead of lasting about 6.5 hours, it was done after 3.5. Even more impressive was that while the mapping time only went down by about 15%, the reducing time was slashed from 3 hours to 30 minutes.

This is the jobs cluster CPU Usage and Load Average after we made the changes

The Cluster Utilization shows a very clear picture. The overall utilization and load average during mapping phase actually increased a bit and instead of lasting 3 hours 10 minutes it was done after 2 hours and 40 minutes. While not huge this is still a 15% improvement.

The reduce phase on the other hand shrank dramatically: from roughly 3 hours to 30 minutes. That means a couple of hours of development work lead to an impressive sixfold performance improvement. We also see that the reduce phase is of course still not utilizing the whole cluster and it's actually the 100% phase that got a lot shorter.

Conclusion
Three simple code fixes resulted in a 100% improvement of our biggest job and a sixfold speedup of the reduce portion. Suffice it to say that this totally surprised the owners of the job. The job was utilizing 100% of the cluster, which for them meant that from a Hadoop perspective things were running in an optimal fashion. While this is true, it doesn't mean that the job itself is efficient.

This example shows that optimizing MapReduce jobs beyond tinkering with Hadoop options can lead to a lot more efficiency without adding any more hardware - achieving the same result with fewer resources.

The Hotspot analysis did also reveal some Hadoop-specific hotspots that led us to change some job options and speed things up even more. More on that in my next blog.

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 (0)

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.


@CloudExpo Stories
"ZeroStack is a startup in Silicon Valley. We're solving a very interesting problem around bringing public cloud convenience with private cloud control for enterprises and mid-size companies," explained Kamesh Pemmaraju, VP of Product Management at ZeroStack, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
In his session at 21st Cloud Expo, Carl J. Levine, Senior Technical Evangelist for NS1, will objectively discuss how DNS is used to solve Digital Transformation challenges in large SaaS applications, CDNs, AdTech platforms, and other demanding use cases. Carl J. Levine is the Senior Technical Evangelist for NS1. A veteran of the Internet Infrastructure space, he has over a decade of experience with startups, networking protocols and Internet infrastructure, combined with the unique ability to it...
"Codigm is based on the cloud and we are here to explore marketing opportunities in America. Our mission is to make an ecosystem of the SW environment that anyone can understand, learn, teach, and develop the SW on the cloud," explained Sung Tae Ryu, CEO of Codigm, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
High-velocity engineering teams are applying not only continuous delivery processes, but also lessons in experimentation from established leaders like Amazon, Netflix, and Facebook. These companies have made experimentation a foundation for their release processes, allowing them to try out major feature releases and redesigns within smaller groups before making them broadly available. In his session at 21st Cloud Expo, Brian Lucas, Senior Staff Engineer at Optimizely, discussed how by using ne...
"There's plenty of bandwidth out there but it's never in the right place. So what Cedexis does is uses data to work out the best pathways to get data from the origin to the person who wants to get it," explained Simon Jones, Evangelist and Head of Marketing at Cedexis, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
"Cloud Academy is an enterprise training platform for the cloud, specifically public clouds. We offer guided learning experiences on AWS, Azure, Google Cloud and all the surrounding methodologies and technologies that you need to know and your teams need to know in order to leverage the full benefits of the cloud," explained Alex Brower, VP of Marketing at Cloud Academy, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clar...
Large industrial manufacturing organizations are adopting the agile principles of cloud software companies. The industrial manufacturing development process has not scaled over time. Now that design CAD teams are geographically distributed, centralizing their work is key. With large multi-gigabyte projects, outdated tools have stifled industrial team agility, time-to-market milestones, and impacted P&L stakeholders.
Gemini is Yahoo’s native and search advertising platform. To ensure the quality of a complex distributed system that spans multiple products and components and across various desktop websites and mobile app and web experiences – both Yahoo owned and operated and third-party syndication (supply), with complex interaction with more than a billion users and numerous advertisers globally (demand) – it becomes imperative to automate a set of end-to-end tests 24x7 to detect bugs and regression. In th...
Enterprises are moving to the cloud faster than most of us in security expected. CIOs are going from 0 to 100 in cloud adoption and leaving security teams in the dust. Once cloud is part of an enterprise stack, it’s unclear who has responsibility for the protection of applications, services, and data. When cloud breaches occur, whether active compromise or a publicly accessible database, the blame must fall on both service providers and users. In his session at 21st Cloud Expo, Ben Johnson, C...
"Infoblox does DNS, DHCP and IP address management for not only enterprise networks but cloud networks as well. Customers are looking for a single platform that can extend not only in their private enterprise environment but private cloud, public cloud, tracking all the IP space and everything that is going on in that environment," explained Steve Salo, Principal Systems Engineer at Infoblox, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Conventio...
Data scientists must access high-performance computing resources across a wide-area network. To achieve cloud-based HPC visualization, researchers must transfer datasets and visualization results efficiently. HPC clusters now compute GPU-accelerated visualization in the cloud cluster. To efficiently display results remotely, a high-performance, low-latency protocol transfers the display from the cluster to a remote desktop. Further, tools to easily mount remote datasets and efficiently transfer...
"MobiDev is a software development company and we do complex, custom software development for everybody from entrepreneurs to large enterprises," explained Alan Winters, U.S. Head of Business Development at MobiDev, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
"We're developing a software that is based on the cloud environment and we are providing those services to corporations and the general public," explained Seungmin Kim, CEO/CTO of SM Systems Inc., in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
Agile has finally jumped the technology shark, expanding outside the software world. Enterprises are now increasingly adopting Agile practices across their organizations in order to successfully navigate the disruptive waters that threaten to drown them. In our quest for establishing change as a core competency in our organizations, this business-centric notion of Agile is an essential component of Agile Digital Transformation. In the years since the publication of the Agile Manifesto, the conn...
SYS-CON Events announced today that CrowdReviews.com has been named “Media Sponsor” of SYS-CON's 22nd International Cloud Expo, which will take place on June 5–7, 2018, at the Javits Center in New York City, NY. CrowdReviews.com is a transparent online platform for determining which products and services are the best based on the opinion of the crowd. The crowd consists of Internet users that have experienced products and services first-hand and have an interest in letting other potential buye...
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, provided 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 oper...
"IBM is really all in on blockchain. We take a look at sort of the history of blockchain ledger technologies. It started out with bitcoin, Ethereum, and IBM evaluated these particular blockchain technologies and found they were anonymous and permissionless and that many companies were looking for permissioned blockchain," stated René Bostic, Technical VP of the IBM Cloud Unit in North America, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Conventi...
SYS-CON Events announced today that Telecom Reseller has been named “Media Sponsor” of SYS-CON's 22nd International Cloud Expo, which will take place on June 5-7, 2018, at the Javits Center in New York, NY. Telecom Reseller reports on Unified Communications, UCaaS, BPaaS for enterprise and SMBs. They report extensively on both customer premises based solutions such as IP-PBX as well as cloud based and hosted platforms.
In his session at 21st Cloud Expo, James Henry, Co-CEO/CTO of Calgary Scientific Inc., introduced you to the challenges, solutions and benefits of training AI systems to solve visual problems with an emphasis on improving AIs with continuous training in the field. He explored applications in several industries and discussed technologies that allow the deployment of advanced visualization solutions to the cloud.
While some developers care passionately about how data centers and clouds are architected, for most, it is only the end result that matters. To the majority of companies, technology exists to solve a business problem, and only delivers value when it is solving that problem. 2017 brings the mainstream adoption of containers for production workloads. In his session at 21st Cloud Expo, Ben McCormack, VP of Operations at Evernote, discussed how data centers of the future will be managed, how the p...