Working with big data and open source software
I recently sat down with Mark Grover (@mark_grover), a Software Engineer at Cloudera, to talk about the Hadoop ecosystem. He is a committer on Apache Bigtop and a contributor to Apache Hadoop, Hive, Sqoop, and Flume. He also contributed to O’Reilly Media’s Programming Hive title.
Key highlights include:
Moving different workloads and frameworks onto the same collection of machines increases efficiency and ROI
As organizations increasingly rely on large computing clusters, tools for leveraging and efficiently managing compute resources become critical. Specifically, tools that allow multiple services and frameworks run on the same cluster can significantly increase utilization and efficiency. Schedulers1 take into account policies and workloads to match jobs with appropriate resources (e.g., memory, storage, processing power) in a large compute cluster. With the help of schedulers, end users begin thinking of a large cluster as a single resource (like “a laptop”) that can be used to run different frameworks (e.g., Spark, Storm, Ruby on Rails, etc.).
Multi-tenancy and efficient utilization translates into improved ROI. Google’s scheduler, Borg, has been in production for many years and has led to substantial savings2. The company’s clusters handle a variety of workloads that can be roughly grouped into batch (compute something, then finish) and services (web or infrastructure services like BigTable). Researchers recently examined traces from several Google clusters and observed that while “batch jobs” accounted for 80% of all jobs, “long service jobs” utilize 55-60% of resources.
There are other benefits of multi-tenancy. Being able to run analytics (batch, streaming) and long running services (e.g., web applications) on the same cluster significantly lowers latency3, opening up the possibility for real-time, analytic applications. Bake-offs can be done more effectively as competing tools, versions, and frameworks can be deployed on the same cluster. Data scientists and production engineers leverage the same compute resources, making it easier for teams to work together across the analytic lifecycle. An additional benefit is that data science teams learn to build products and services that factor in efficient utilization and availability.
Mesos, Chronos, and Marathon
Apache Mesos is a popular open source scheduler that originated from UC Berkeley’s AMPlab. Mesos is based on features in modern kernels for resource isolation (cgroups in Linux). It has been in production for a few years at Twitter4, airbnb5, and many other companies – AMPlab simulations showed Mesos comfortably handling clusters with 30K servers.
As data sizes continue to grow, interactive query systems may start adopting the sampling approach central to BlinkDB
Interactive query analysis for (Hadoop scale data) has recently attracted the attention of many companies and open source developers – some examples include Cloudera’s Impala, Shark, Pivotal’s HAWQ, Hadapt, CitusDB, Phoenix, Sqrrl, Redshift, and BigQuery. These solutions use distributed computing, and a combination of other techniques including data co-partitioning, caching (into main memory), runtime code generation, and columnar storage.
One approach that hasn’t been exploited as much is sampling. By this I mean employing samples to generate approximate answers, and speed up execution. Database researchers have written papers on approximate answers, but few working (downloadable) systems are actually built on this approach.
Approximate query engine from U.C. Berkeley’s Amplab
An interesting, open source database released yesterday0 uses sampling to scale to big data. BlinkDB is a massively-parallel, approximate query system from UC Berkeley’s Amplab. It uses a series of data samples to generate approximate answers. Users compose queries by specifying either error bounds or time constraints, BlinkDB uses sufficiently large random samples to produce answers. Because random samples are stored in memory1, BlinkDB is able to provide interactive response times:
Compelling large-scale data platforms originate from the world of IT Operations
I’ve been noticing that many interesting big data systems are coming out of IT operations. These are systems that go beyond the standard “capture/measure, display charts, and send alerts”. IT operations has long been a source of many interesting big data1 problems and I love that it’s beginning to attract the attention2 of many more data scientists and data engineers.
It’s not surprising that many of the interesting large-scale systems that target time-series and event data have come from ops teams: in an earlier post on time-series, several of the tools I highlighted came out of IT operations. IT operations involves monitoring many different hardware and software systems, a task that requires a variety of tools and which quickly leads to “metrics overload”. A partial list includes data captured from a wide range of application log files, network traffic, energy and power sources.
The volume of IT ops data has led to new tools like OpenTSDB and KairosDB – time series databases that leverage HBase and Cassandra. But storage, simple charts, and lookups are just the foundation of what’s needed. IT Ops track many interdependent systems, some of which might be correlated3. Not only are IT ops faced with highlighting “unknown unknowns” in their massive data sets, they often need to do so in near realtime.
Eleven areas of focus for deeper investigation.
Conferences like Strata are planned a year in advance. The logistics and coordination required for an event of this magnitude takes a lot of planning, but it also takes a decent amount of prediction: Strata needs to skate to where the puck is going.
While Strata New York + Hadoop World 2013 is still a few months away, we’re already guessing at what next year’s Santa Clara event will hold. Recently, the team got together to identify some of the hot topics in big data, ubiquitous computing, and new interfaces. We selected eleven big topics for deeper investigation.
- Deep learning
- Time-series data
- The big data “app stack”
- Cultural barriers to change
- Design patterns
- Laggards and Luddites
- The convergence of two databases
- The other stacks
- Mobile data
- The analytic life-cycle
- Data anthropology
Here’s a bit more detail on each of them. Read more…
Collaborative filtering with Neo4j
By this time, chances are very likely that you’ve heard of NoSQL, and of graph databases like Neo4j.
NoSQL databases address important challenges that we face today, in terms of data size and data complexity. They offer a valuable solution by providing particular data models to address these dimensions.
On one side of the spectrum, these databases resolve issues for scaling out and high data values using compounded aggregate values, on the other side is a relationship based data model that allows us to model real world information containing high fidelity and complexity.
Neo4j, like many other graph databases, builds upon the property graph model; labeled nodes (for informational entities) are connected via directed, typed relationships. Both nodes and relationships hold arbitrary properties (key-value pairs). There is no rigid schema, but with node-labels and relationship-types we can have as much meta-information as we like. When importing data into a graph database, the relationships are treated with as much value as the database records themselves. This allows the engine to navigate your connections between nodes in constant time. That compares favorably to the exponential slowdown of many-JOIN SQL-queries in a relational database.
How can you use a graph database?
Graph databases are well suited to model rich domains. Both object models and ER-diagrams are already graphs and provide a hint at the whiteboard-friendliness of the data model and the low-friction mapping of objects into graphs.
Instead of de-normalizing for performance, you would normalize interesting attributes into their own nodes, making it much easier to move, filter and aggregate along these lines. Content and asset management, job-finding, recommendations based on weighted relationships to relevant attribute-nodes are some use cases that fit this model very well.
Many people use graph databases because of their high performance online query capabilities. They process large amounts or high volumes of raw data with Map/Reduce in Hadoop or Event-Processing (like Storm, Esper, etc.) and project the computation results into a graph. We’ve seen examples of this from many domains from financial (fraud detection in money flow graphs), biotech (protein analysis on genome sequencing data) to telco (mobile network optimizations on signal-strength-measurements).
Graph databases shine when you can express your queries as a local search using a few starting points (e.g., people, products, places, orders). From there, you can follow relevant relationships to accumulate interesting information, or project visited nodes and relationships into a suitable result.
A new set of analytic engines make the case for convenience over performance
The choice of tools for data science includes1 factors like scalability, performance, and convenience. A while back I noted that data scientists tended to fall into two camps: those who used an integrated stack, and others who tended to stitch together frameworks. Being able to stick with the same programming language and environment is a definite productivity boost since it requires less setup time and context-switching.
More recently I highlighted the emergence of composable analytic engines, that leverage data stored in HDFS (or HBase and Accumulo). These engines may not be the fastest available, but they scale to data sizes that cover most workloads, and most importantly they can operate on data stored in popular distributed data stores. The fastest and most complete set of algorithms will still come in handy, but I suspect that users will opt for slightly slower2, but more convenient tools, for many routine analytic tasks.
OSCON 2013 Speaker Series
If you have delved into Apache Hadoop and related projects, you know that installing and configuring Hadoop is hard. Often, a minor mistake during installation or configuration with messy tarballs will lurk for a long time until some otherwise innocuous change to the system or workload causes difficulties. Moreover, there is little to no integration testing among different projects (e.g. Hadoop, Hive, HBase, Zookeeper, etc.) in the ecosystem. Apache Bigtop is an open source project aimed at bridging exactly those gaps by:
1. Making it easier for users to deploy and configure Hadoop and related projects on their bare metal or virtualized clusters.
2. Performing integration testing among various components in the Hadoop ecosystem.
More about Apache Bigtop
The primary goal of Apache Bigtop is to build a community around the packaging and interoperability testing of Hadoop related projects. This includes testing at various levels (packaging, platform, runtime, upgrade, etc.) developed by a community with a focus on the system as a whole, rather than individual projects.
The latest released version of Apache Bigtop is Bigtop 0.5 which integrates the latest versions of various projects including Hadoop, Hive, HBase, Flume, Sqoop, Oozie and many more! The supported platforms include CentOS/RHEL 5 and 6, Fedora 16 and 17, SuSE Linux Enterprise 11, OpenSuSE 12.2, Ubuntu LTS Lucid and Precise, and Ubuntu Quantal.
Who uses Bigtop?
Folks who use Bigtop can be divided into two major categories. The first category of users are those who leverage Bigtop to power their own Hadoop Distributions. The second category of users are those who use Bigtop for deployment purposes.
In alphabetical order, they are:
OSCON 2013 Speaker Series
Paco Nathan (@pacoid) is Director of Data Science at Concurrent, O’Reilly Author, and OSCON 2013 Speaker. In this interview we talk about creating enterprise data workflow with Cascading. Be sure to check out Paco’s book on the subject here
NOTE: If you are interested in attending OSCON to check out Paco’s talk or the many other cool sessions, click over to the OSCON website where you can use the discount code OS13PROG to get 20% your registration fee.
Key highlights include:
- Cascading is an abstraction layer on top of Hadoop [Discussed at 0:23]
- Define your business logic at a high level [Discussed at 1:21]
- Is Cascading good for enterprise? [Discussed at 2:31]
- Test-driven development at scale [Discussed at 3:35]
- Cascalog and the City of Palo Alto Open Data portal [Discussed at 7:39]
You can view the full interview here:
OSCON 2013 Speaker Series
NOTE: If you are interested in attending OSCON to check out Dave’s talk or the many other cool sessions, click over to the OSCON website where you can use the discount code OS13PROG to get 20% off your registration fee.
Since 2009, I’ve been leading the optimization team at AppNexus, a real-time advertising exchange. On this exchange, advertisers participate in real-time auctions to bid on individual ad impressions. The highest bid wins the auction, and that advertiser gets to show an ad. This allows advertisers to carefully target where they advertise—maximizing the effectiveness of their advertising budget—and lets websites maximize their ad revenue.
We do these auctions often (~50 billion a day) and fast (<100 milliseconds). Not surprisingly, this creates a lot of technical challenges. One of those challenges is how to automatically maximize the value advertisers get for their marketing budgets—systematically driving consumer engagement through ad placements on particular websites, times of day, etc.—and we call this process “optimization.” The volume of data is large, and the algorithms and strategies aren’t trivial.
In order to win clients and build our business to the scale we have today, it was crucial that we build a world-class optimization system. But when I started, we didn’t have a scalable tech stack to process the terabytes of data flowing through our systems every day, and we didn't have the team to do any of the required data modeling.
So, we needed to hire great people fast. However, there aren’t many veterans in the advertising optimization space, and because of that, we couldn’t afford to narrow our search to only experts in Java or R or Matlab. In order to give us the largest talent pool possible to recruit from, we had to choose a tech stack that is both powerful and accessible to people with diverse experience and backgrounds. So we chose Python.
Python is easy to learn. We found that people coding in R, Matlab, Java, PHP, and even those who have never programmed before could quickly learn and get up to speed with Python. This opened us up to hiring a tremendous pool of talent who we could train in Python once they joined AppNexus. To top it off, there’s a great community for hiring engineers and the PyData community is full of programmers who specialize in modeling and automation.
Additionally, Python has great libraries for data modeling. It offers great analytical tools for analysts and quants and when combined, Pandas, IPython, and Matplotlib give you a lot of the functionality of Matlab or R. This made it easy to hire and onboard our quants and analysts who were familiar with those technologies. Even better, analysts and quants can share their analysis through the browser with IPython.
Now that we had all of these wonderful employees, we needed a way to cut down the time to get them ramped up and pushing code to production.
First, we wanted to get our analysts and quants looking at and modeling data as soon as possible. We didn’t want them worrying about writing database connector code, or figuring out how to turn a cursor into a data frame. To tackle this, we built a project called Link.
Imagine you have a MySQL database. You don’t want to hardcode all of your connection information because you want to have a different config for different users, or for different environments. Link allows you to define your “environment” in a JSON config file, and then reference it in code as if it is a Python object.
Now, with only three lines of code you have a database connection and a data frame straight from your mysql database. This same methodology works for Vertica, Netezza, Postgres, Sqlite, etc. New “wrappers” can be added to accommodate new technologies, allowing team members to focus on modeling the data, not how to connect to all these weird data sources.
In : from link import lnk
In : my_db = lnk.dbs.my_db
In : df = my_db.select('select * from my_table').as_dataframe()
Int64Index: 325 entries, 0 to 324
id 325 non-null values
user_id 323 non-null values
app_id 325 non-null values
name 325 non-null values
body 325 non-null values
created 324 non-null values
By having the flexibility to easily connect to new data sources and APIs, our quants were able to adapt to the evolving architectures around us, and stay focused on modeling data and creating algorithms.
Second, we wanted to minimize the amount of work it took to take an algorithm from research/prototype phase to full production scale. Luckily, with everyone working in Python, our quants, analysts, and engineers are using the same language and data processing libraries. There was no need to re-implement an R script in Java to get it out across the platform.