We are building a real-time big data tool with open source tools. Our main goal is to supervise and analyze a network by getting logs from a kafka server in real-time. We saw in tutorials that we have to divide our tool in two sections: Analytic and Supervision as shown below.
For the supervision section we chose the solution Elasticsearch and Logstash.
Regarding the section analytic, my team and I are comparing Apache Storm Streaming and Apache Storm in order to use it with Elasticsearch. Despite the fact that Apache Storm is a true real-time data processing tool and faster than Apache Spark Streaming, it does not provide machine learning libraries like with Apache Spark. That's why we are thinking to choose Apache Spark. The elastic website indicates that it exists a connector ES-Hadoop to connect a Elasticsearch database to a Hadoop ecosystem. We can see that in the below figure.
However, We are a little bit confused with this picture because there is only spark SQL and not all the spark frameworks (MLlib, Spark Streaming..). We did some assumptions and we came out with two final possible architectures. We only wanted to know if there are technically correct and if we are not in the wrong direction.
With Apache Spark streaming:
With Apache Storm:
Both your architectural diagrams are ok. Keep on mind that spark streaming will not work in this scenario. Es-hadoop provides you with easy access apis to get and put data from and into elastic. Its also provides the methods to get the data inro the spark framework (RDD) or data frames inthe case of spark sql. Once the data is in the framework, all ml libraries can be applied to the data for ml or analytics generation. Elastic is not capable of streaming data so spark streaming in the strict sense is not possible. So in the diagram, the arrow to hdfs optional and then to spark streaming can be removed and the arrow juat pointa to hdfs. My concern, however, would be running mllib algos on the data in realtime and expect realtime performance. Typical use case might be do modwl generation off line and use the model in realtime for analysis.
Related
I am working on building a distributed real time cluster system to supervise and analyze a network. I did several researches on internet and I came out with few technologies:
for real time processing : logstash, storm and apache streaming
for storage: elasticsearch
for analysis: Apache Spark over Hadoop (I will use ES-Hadoop to connect with Elasticsearch)
for data visualization: kibana, D3js, c3js
However, logstash is not often mentioned as spark streaming and storm. I found in internet the following architecture presented in the below picture:
I have two questions:
I don't understand why logstash is not often mentioned as a real-tim processing system like spark streaming and storm. What are the main reasons ? I hav been using it and it is very powerful..
Regarding the Analyze part, can I use the machine learning librairies in that configuration ?
Logstash is not cluster stream processing system. It is simply a JVM based process. The latest version supports on disk buffer but does not have the nearly the same delivery guaranties as Spark or Storm. Take a look at http://storm.apache.org/releases/1.0.3/Guaranteeing-message-processing.html
Yes but not sure why use Elastic for storing data first. Why not HDFS->SparkML->Elastic? The main thing to think here is managing models, training and testing.
My understanding was that Spark is an alternative to Hadoop. However, when trying to install Spark, the installation page asks for an existing Hadoop installation. I'm not able to find anything that clarifies that relationship.
Secondly, Spark apparently has good connectivity to Cassandra and Hive. Both have sql style interface. However, Spark has its own sql. Why would one use Cassandra/Hive instead of Spark's native sql? Assuming that this is a brand new project with no existing installation?
Spark is a distributed in memory processing engine. It does not need to be paired with Hadoop, but since Hadoop is one of the most popular big data processing tools, Spark is designed to work well in that environment. For example, Hadoop uses the HDFS (Hadoop Distributed File System) to store its data, so Spark is able to read data from HDFS, and to save results in HDFS.
For speed, Spark keeps its data sets in memory. It will typically start a job by loading data from durable storage, such as HDFS, Hbase, a Cassandra database, etc. Once loaded into memory, Spark can run many transformations on the data set to calculate a desired result. The final result is then typically written back to durable storage.
In terms of it being an alternative to Hadoop, it can be much faster than Hadoop at certain operations. For example a multi-pass map reduce operation can be dramatically faster in Spark than with Hadoop map reduce since most of the disk I/O of Hadoop is avoided. Spark can read data formatted for Apache Hive, so Spark SQL can be much faster than using HQL (Hive Query Language).
Cassandra has its own native query language called CQL (Cassandra Query Language), but it is a small subset of full SQL and is quite poor for things like aggregation and ad hoc queries. So when Spark is paired with Cassandra, it offers a more feature rich query language and allows you to do data analytics that native CQL doesn't provide.
Another use case for Spark is for stream processing. Spark can be set up to ingest incoming real time data and process it in micro-batches, and then save the result to durable storage, such as HDFS, Cassandra, etc.
So spark is really a standalone in memory system that can be paired with many different distributed databases and file systems to add performance, a more complete SQL implementation, and features they may lack such a stream processing.
Im writing a paper about Hadoop for university. And stumbled over your question. Spark is just using Hadoop for persistence and only if you want to use it. It's possible to use it with other persistence tiers like Amazon EC2.
On the other hand-side spark is running in-memory and it's not primarly build to be used for map reduce use-cases like Hadoop was/is.
I can recommend this article, if you like a more detailed description: https://www.xplenty.com/blog/2014/11/apache-spark-vs-hadoop-mapreduce/
The README.md file in Spark can solve your puzzle:
A Note About Hadoop Versions
Spark uses the Hadoop core library to talk to HDFS and other Hadoop-supported
storage systems. Because the protocols have changed in different versions of
Hadoop, you must build Spark against the same version that your cluster runs.
Please refer to the build documentation at
"Specifying the Hadoop Version"
for detailed guidance on building for a particular distribution of Hadoop, including
building for particular Hive and Hive Thriftserver distributions.
I am new to all these terms and given some time to understand it. But i have some confusions in it. Please correct me if i am wrong.
Nutch: It's for web crawling, using it we can crawl web pages. We can store these web pages somewhere in db.
Solr: Solr can be used for indexing web pages crawled by Apache Nutch. It helps in searching the indexes web pages.
HBase: It's used as an interface to interact with Hadoop. It helps in getting data at real time from HDFS. It provides simple SQL type interface for interacting.
Hadoop: It provides two functionalities: One is HDFS (Hadoop data file system) and other is Map-Reduce functionality taken from Google algorithms. Its basically used for offline data backup etc.
Gora and ZooKeeper: I am not sure of.
Confusions:
1). Is HBase a key-value pair DB or just an interface to Hadoop ? or i should ask, can HBase exist without Hadoop ?
If yes, can you explain a bit more about its usage.
2). Is there any use of crawling data using Apache Nutch without indexing into Solr ?
3). For running apache nutch, do we need HBase and Hadoop ? If no, how we can make it work without it?
4). Is Hadoop part of HBase ?
Here is a good short discussion of HBase vs. Hadoop: Difference between HBase and Hadoop/HDFS
Because HBase is built on top of Hadoop you can't really have HBase without Hadoop.
Yes you can run Nutch without Solr; there do not seem to be lots of use cases, however, much less living examples in the wild.
Yes, you can run Nutch without Hadoop, but again there don't seem to be a lot of real-world examples of people doing this.
Yes Hadoop is part of HBase, in that there is no HBase without Hadoop, but of course Hadoop is used for other things as well.
Zookeeper is used for configuration, naming, synchronization, etc. in Hadoop stack workflows. Gora is a memory management/persistence framework and is built on top of Hadoop.
I have just started exploring BigData technology and the Hadoop framework.
But, getting confused with so many ecosystem components and framework. Could you please advise to get a structured start for learning ?
I mean which ecosystem component should one focus? Any in particular or all?
Help much appreciated!
Ranit
I wrote this answer on Quora few months back. Hope this will help:
1. Go through some introductory videos on Hadoop
Its very important to have some high level idea of hadoop before directly starting working on it. These introductory videos will help in understanding the scope of Hadoop and the use cases where it can be applied. There are a lot of resources available online for the same and going through any of the videos will be beneficial.
2. Understanding MapReduce
The second thing which helped me was to understand what Map Reduce is and how it works. It is explained very nicely in this paper: http://static.googleusercontent....
Another nice tutorial is available here : http://ksat.me/map-reduce-a-real...
For points 1 and 2, go through first four lectures for week one video lectures. The whole concept of distributed computing and map reduce is explained very nicely here. https://class.coursera.org/mmds-001/lecture
3. Getting started with Cloudera VM
Once you understand the basics of Hadoop, you can download the VM provided by cloudera and starting running some hadoop commands on it. You can download the VM from this link: http://www.cloudera.com/content/...
It would be nice to get familiar with basic Hadoop commands on the VM and understanding how it works.
4. Setting up the standalone/Pseudo distributed Hadoop
I would recommend setting up your own standalone Hadoop on your machine once you are familiar with Hadoop using the VM. The steps for installing are explained very nicely on this blog by Michael G. Noll : Running Hadoop On Ubuntu Linux (Single-Node Cluster) - Michael G. Noll
5. Understanding the Hadoop Ecosystem
It would be nice to get familiar with other components in the Hadoop ecosystem like Apache Pig, Hive, Hbase, Flume-NG, Hue etc. All these serve different purposes and having some information on all these will be really helpful in building any product around the hadoop ecosystem. You can install all these easily on your machine and get started with them. Cloudera VM by has most of these installed already.
6. Writing Map Reduce Jobs
Once you are done with steps 1-5, I don't think writing Map Reduce would be a challenge. It is explained thoroughly in The Definitive Guide. If MapReduce really interests you a lot, I would suggest reading this book Mining Massive Datasets by Anand Rajaraman, Jure Leskovec and Jeffrey D. Ullman : Page on Stanford
I would recommend going for Hadoop first, it's the basis for a lot of those other systems out there. Check out the main site: http://hadoop.apache.org/ and check out Cloudera, they provide a Virtual image (called CDH), that comes with everything pre-installed, so you can jump into action without having to deal with installation problems: http://www.cloudera.com/content/cloudera/en/downloads/cdh/cdh-5-2-0.html
After that, I would look into HDFS, just to understand a bit more how Hadoop stores that data, and then it would depend on what type of problems you're trying to solve, each particular system tackles a specific and (usually) different problem:
Hive / Cassandra: For database-like interaction
Pig: For data transformation.
Spark: For real time data analysis
Check out this link for more details: http://www.cloudera.com/content/cloudera/en/training/library/apache-hadoop-ecosystem.html
I hope you find that useful.
Big data is a broad term for data sets so large or complex that traditional data processing applications are inadequate. Challenges include analysis, capture, data curation, search, sharing, storage, transfer, visualization, and information privacy - From wikipedia
Hadoop is a a framework that allows for the distributed processing of large data sets across clusters of computers using simple programming models.
There are four main modules in Hadoop.
1.Hadoop Common: The common utilities that support the other Hadoop modules.
2.Hadoop Distributed File System (HDFS™): A distributed file system that provides high-throughput access to application data.
3.Hadoop YARN: A framework for job scheduling and cluster resource management.
4.Hadoop MapReduce: A YARN-based system for parallel processing of large data sets.
Before going further, Let's note that we have three different types of data.
Structured : Structured data has strong schema and schema will be checked during write & read operation. e.g. Data in RDBMS systems like Oracle, MySQL Server etc.
Unstructured: Data does not have any structure and it can be any form - Web server logs, E-Mail, Images etc.
Semi-structured: Data is not strictly structured but have some structure. e.g. XML files
Depending on type of data to be processed, we have to choose right technology.
Some more projects, which are part of Hadoop
HBase™: A scalable, distributed database that supports structured data storage for large tables.
Hive™: A data warehouse infrastructure that provides data summarization and ad hoc querying.
Pig™: A high-level data-flow language and execution framework for parallel computation
Hive Vs PIG comparison can be found at my other post in this question
HBASE won't replace Map Reduce. HBase is scalable distributed database & Map Reduce is programming model for distributed processing of data. Map Reduce may act on data in HBASE in processing.
You can use HIVE/HBASE for structured/semi-structured data and process it with Hadoop Map Reduce
You can use SQOOP to import structured data from traditional RDBMS database Oracle, SQL Server etc and process it with Hadoop Map Reduce
You can use FLUME for processing Un-structured data and process with Hadoop Map Reduce
Have a look at: Hadoop Use Cases
Hive should be used for analytical querying of data collected over a period of time. e.g Calculate trends , summarize website logs but it can't be used for real time queries.
HBase fits for real-time querying of Big Data. Facebook use it for messaging and real-time analytics.
PIG can be used to construct dataflows,run a scheduled jobs, crunch big volumes of data,aggregate/summarize it and store into relation database systems. Good for ad-hoc analysis.
Hive can be used for ad-hoc data analysis but it can't support all un-structured data formats unlike PIG
ZooKeeper is a centralized service for maintaining configuration information, naming, providing distributed synchronization and providing group services which are very useful for a variety of distributed systems. HBase is not operational without ZooKeeper.
Apache Spark is a general compute engine that offers fast data analysis on a large scale. Spark is built on HDFS but bypasses MapReduce and instead uses its own data processing framework. Common uses cases for Apache Spark include real-time queries, event stream processing, iterative algorithms, complex operations and machine learning.
Mahout™: A Scalable machine learning and data mining library.
Tez™: A generalized data-flow programming framework, built on Hadoop YARN, which provides a powerful and flexible engine to execute an arbitrary DAG of tasks to process data for both batch and interactive use-cases. Tez is being adopted by Hive™, Pig™ and other frameworks in the Hadoop ecosystem, and also by other commercial software (e.g. ETL tools), to replace Hadoop™ MapReduce as the underlying execution engine
I have covered only some of key components of Hadoop ecosystem. If you like to have a look at all component of ecosystem, have a look at this ecosystem table
If above table is very difficult to digest, have a look at minified version of ecosystem at this article
But to understand all of these system, I would like you to start with Apache website first and explore other articles later.
Big data is not a technology in itself, instead it is a concept.
You can think of database, database is not a technology in itself, it is a concept. Oracle, DB2 etc are database technologies.
So coming back to big data, this concept is used to deal with huge data which is difficult to be analyzed using traditional databases or technologies. People think hadoop as synonym of bigdata but again let me tell you that Hadoop is nothing but a technology developed by Apache to implement bigdata concept.
Hadoop has its own file system called hdfs and it uses mapreduce to solve bigdata problems. Apart from Hadoop there is hive which is similar to sql but internally it uses map reduce. Hbase is similar to nosql database. Pig is scripting language which uses mapreduce internally.
There are many licensed version for big data like MapR, Hortonworks, Cloudera etc.
So start learning with Hadoop - HDFS, Mapreduce, Yarn, Hive.
Things I did to learn Hadoop.
a) Install Hadoop from scratch. I mean download CentOs, Hadoop , JAVA etc., and install them manually.
b) Understand how HDFS works.
c) Understand how MapReduce works.
d) Write word count in JAVA.
This will help you get started.
I appreciate ideas on how to stream data from an On-Premise Windows server to a persistent EMR cluster?
Some Background
I would like to run a persistent cluster running a MR job much like the WordCount examples that are available. I would like to stream text from a local Windows Server up to the cluster and have it processed by the running job.
All of the streaming WordCount examples I have reviewed always start with a static text file in S3 and don't cover how to implement anything to generate the stream.
Does this need to be treated in two parts?
Get the data first into S3
Stream it into the EMR cluster?
I have seen tools like Logstash which tend to run agents on the local server which tail the end of a weblog and transfer it.
As you can probably tell, I'm a Windows guy, stretching into EMR and by association Linux. Feel free to let me know if there is some way cool command line tool that already does this.
Thanks in advance.
Currently EMR as-is only supports MR, Hive, Pig, HBase and Impala. MR/Hive/Pig process the data in a batch oriented fashion and data can't be streamed to them. While HBase is a NoSQL DB and Impala is used for interactive ad-hoc queries.
For processing streaming data there are a lot of other options like Storm, Samza, S4. From AWS there is Kinesis which has been moved into GA recently.
Yes a static file would go into S3 and then be the input into your EMR cluster job.
But I believe that fact you want a persistent cluster implies you are streaming in continuation from your Windows server. Is that the case ?
If so you need to create a AWS Kinesis Stream, configure your producers which put data into the stream's shards by calling the Putrecord.
Start by reading "Developing Record Consumer Applications"
I think you could use apache Flume (https://flume.apache.org/)
Flume is a distributed, reliable, and available service for efficiently collecting, aggregating, and moving large amounts of log data. It has a simple and flexible architecture based on streaming data flows. It is robust and fault tolerant with tunable reliability mechanisms and many failover and recovery mechanisms. It uses a simple extensible data model that allows for online analytic application.