Big data – part 3

Linkedin maps data visualization by luc legay (cc) (from Flickr)
Linkedin maps data visualization by luc legay (cc) (from Flickr)

I have renamed this series to “Big data” because it’s no longer just about Hadoop (see Hadoop – part 1 & Hadoop – part 2 posts).

To try to partition this space just a bit, there is unstructured data analysis and structured data analysis. Hadoop is used to analyze un-structured data (although Hadoop is used to parse and structure the data).

On the other hand, for structured data there are a number of other options currently available. Namely:

  • EMC Greenplum – a relational database that is available in a software only as well as now as a hardware appliance. Greenplum supports both row or column oriented data structuring and has support for policy based data placement across multiple storage tiers. There is a packaged solution that consists of Greenplum software and a Hadoop distribution running on a GreenPlum appliance.
  • HP Vertica – a column oriented, relational database that is available currently in a software only distribution. Vertica supports aggressive data compression and provides high throughput query performance. They were early supporters of Hadoop integration providing Hadoop MapReduce and Pig API connectors to provide Hadoop access to data in Vertica databases and job scheduling integration.
  • IBM Netezza – a relational database system that is based on proprietary hardware analysis engine configured in a blade system. Netezza is the second oldest solution on this list (see Teradata for the oldest). Since the acquisition by IBM, Netezza now provides their highest performing solution on IBM blade hardware but all of their systems depend on purpose built, FPGA chips designed to perform high speed queries across relational data. Netezza has a number of partners and/or homegrown solutions that provide specialized analysis for specific verticals such as retail, telcom, finserv, and others. Also, Netezza provides tight integration with various Oracle functionality but there doesn’t appear to be much direct integration with Hadoop on thier website.
  • ParAccel – a column based, relational database that is available in a software only solution. ParAccel offers a number of storage deployment options including an all in-memory database, DAS database or SSD database. In addition, ParAccel offers a Blended Scan approach providing a two tier database structure with DAS and SAN storage. There appears to be some integration with Hadoop indicating that data stored in HDFS and structured by MapReduce can be loaded and analyzed by ParAccel.
  • Teradata – a relational databases that is based on a proprietary purpose built appliance hardware. Teradata recently came out with an all SSD, solution which provides very high performance for database queries. The company was started in 1979 and has been very successful in retail, telcom and finserv verticals and offer a number of special purpose applications supporting data analysis for these and other verticals. There appears to be some integration with Hadoop but it’s not prominent on their website.

Probably missing a few other solutions but these appear to be the main ones at the moment.

In any case both Hadoop and most of it’s software-only, structured competition are based on a massively parrallelized/share nothing set of linux servers. The two hardware based solutions listed above (Teradata and Netezza) also operate in a massive parallel processing mode to load and analyze data. Such solutions provide scale-out performance at a reasonable cost to support very large databases (PB of data).

Now that EMC owns Greenplum and HP owns Vertica, we are likely to see more appliance based packaging options for both of these offerings. EMC has taken the lead here and have already announced Greenplum specific appliance packages.

—-

One lingering question about these solutions is why don’t customers use current traditional database systems (Oracle, DB2, Postgres, MySQL) to do this analysis. The answer seems to lie in the fact that these traditional solutions are not massively parallelized. Thus, doing this analysis on TB or PB of data would take a too long. Moreover, the cost to support data analysis with traditional database solutions over PB of data would be prohibitive. For these reasons and the fact that compute power has become so cheap nowadays, structured data analytics for large databases has migrated to these special purpose, massively parallelized solutions.

Comments?

Hadoop – part 2

Hadoop Graphic (c) 2011 Silverton Consulting
Hadoop Graphic (c) 2011 Silverton Consulting

(Sorry about the length).

In part 1 we discussed some of Hadoop’s core characteristics with respect to the Hadoop distributed file system (HDFS) and the MapReduce analytics engine. Now in part 2 we promised to discuss some of the other projects that have emerged to make Hadoop and specifically MapReduce even easier to use to analyze unstructured data.

Specifically, we have a set of tools which use Hadoop to construct a database like out of unstructured data.  Namely,

  • Casandra – which maps HDFS data into a database but into a columnar based sparse table structure rather than the more traditional relational database row form. Cassandra was written by Facebook for Mbox search. Columnar databases support a sparse data much more efficiently.  Data access is via a Thrift based API supporting many languages.  Casandra’s data model is based on column, column families and column super-families. The datum for any column item is a three value structure and consists of a name, value of item and a time stamp.  One nice thing about Cassandra is that one can tune it for any consistency model one requires, from no consistency to always consistent and points inbetween.  Also Casandra is optimized for writes.  Cassandra can be used as the Map portion of a MapReduce run.
  • Hbase – which also maps HDFS data into a database like structure and provides Java API access to this DB.  Hbase is useful for million row tables with arbitrary column counts. Apparently Hbase is an outgrowth of Google’s Bigtable which did much the same thing only against the Google file system (GFS).  In contrast to Hive below Hbase doesn’t run on top of MapReduce rather it replaces MapReduce, however it can be used as a source or target of MapReduce operations.  Also, Hbase is somewhat tuned for random access read operations and as such, can be used to support some transaction oriented applications.  Moreover, Hbase can run on HDFS or Amazon S3 infrastructure.
  • Hive – which maps a” simple SQL” (called QL) ontop of a data warehouse built on Hadoop.  Some of these queries may take a long time to execute and as the HDFS data is unstructured the map function must extract the data using a database like schema into something approximating a relational database. Hive operates ontop of Hadoop’s MapReduce function.
  • Hypertable – is a Google open source project which is a  c++ implementation of BigTable only using HDFS rather than GFS .  Actually Hypertable can use any distributed file systemand and is another columnar database (like Cassandra above) but only supports columns and column families.   Hypertable supports both a client (c++) and Thrift API.  Also Hypertable is written in c++ and is considered the most optimized of the Hadoop oriented databases (although there is some debate here).
  • Pig – is a dataflow processing (scripting) language built ontop of Hadoop which supports a sort of database interpreter for HDFS  in combination with an interpretive analysis.  Essentially, Pig uses the scripting language and emits a dataflow graph which is then used by MapReduce to analyze the data in HDFS.  Pig supports both batch and interactive execution but can also be used through a Java API.

Hadoop also supports special purpose tools used for very specialized analysis such as

  • Mahout – an Apache open source project which applies machine learning algorithms to HDFS data providing classification, characterization, and other feature extraction.  However, Mahout works on non-Hadoop clusters as well.  Mahout supports 4 techniques: recommendation mining, clustering, classification, and itemset machine learning functions.  While Mahout uses the MapReduce framework of Hadoop, it doesnot appear that Mahout uses Hadoop MapReduce directly but is rather a replacement for MapReduce focused on machine learning activities.
  • Hama – an Apache open source project which is used to perform paralleled matrix and graph computations against Hadoop cluster data.  The focus here is on scientific computation.  Hama also supports non-Hadoop frameworks including BSP and Dryad (DryadLINQ?). Hama operates ontop of MapReduce and can take advantage of Hbase data structures.

There are other tools that have sprung up around Hadoop to make it easier to configure, test and use, namely

  • Chukwa – which is used for monitoring large distributed clusters of servers.
  • ZooKeeper – which is a cluster configuration tool  and distributed serialization manager useful to build large clusters of Hadoop nodes.
  • MRunit – which is used to unit test MapReduce programs without having to test it on the whole cluster.
  • Whirr – which extends HDFS to use cloud storage services, unclear how well this would work with PBs of data to be processed but maybe it can colocate the data and the compute activities into the same cloud data center.

As for who uses these tools, Facebook uses Hive and Cassandra, Yahoo uses Pig, Google uses Hypertable and there are myriad users of the other projects as well.  In most cases the company identified in the previous list developed the program source code originally, and then contributed it to the Apache for use in the Hadoop open source project. In addition, those companies continue to fix, support and enhance these packages as well.

Hadoop – part 1

Hadoop Logo (from http://hadoop.apache.org website)
Hadoop Logo (from http://hadoop.apache.org website)

BIGData is creating quite a storm around IT these days and at the bottom of big data is an Apache open source project called Hadoop.

In addition, over the last month or so at least three large storage vendors have announced tie-ins with Hadoop, namely EMC (new distribution and product offering), IBM ($100M in research) and NetApp (new product offering).

What is Hadoop and why is it important

Ok, lot’s of money, time and effort are going into deploying and supporting Hadoop on storage vendor product lines. But why Hadoop?

Essentially, Hadoop is a batch processing system for a cluster of nodes that provides the underpinnings of most BIGData analytic activities because it bundle two sets of functionality most needed to deal with large unstructured datasets. Specifically,

  • Distributed file system – Hadoop’s Distributed File System (HDFS) operates using one (or two) meta-data servers (NameNode) with any number of data server nodes (DataNode).  These are essentially software packages running on server hardware which supports file system services using a client file access API. HDFS supports a WORM file system, that splits file data up into segments (64-128MB each), distributes segments to data-nodes, stored on local (or networked) storage and keeps track of everything.  HDFS provides a global name space for its file system, uses replication to insure data availability, and provides widely distributed access to file data.
  • MapReduce processing – Hadoop’s MapReduce functionality is used to parse unstructured files into some sort of structure (map function) which can then be later sorted, analysed and summarized (reduce function) into some worthy output.  MapReduce uses the HDFS to access file segments and to store reduced results.  MapReduce jobs are deployed over a master (JobTracker) and slave (TaskTracker) set of nodes. The JobTracker schedules jobs and allocates activities to TaskTracker nodes which execute the map and reduce processes requested. These activities execute on the same nodes that are used for the HDFS.

Hadoop explained

It just so happens that HDFS is optimized for sequential read access and can handle files that are 100TBs or larger. By throwing more data nodes at the problem, throughput can be scaled up enormously so that a 100TB file could literally be read in minutes to hours rather than weeks.

Similarly, MapReduce parcels out the processing steps required to analyze this massive amount of  data onto these same DataNodes, thus distributing the processing of the data, reducing time to process this data to minutes.

HDFS can also be “rack aware” and as such, will try to allocate file segment replicas to different racks where possible.  In addition, replication can be defined on a file basis and normally uses 3 replicas of each file segment.

Another characteristic of Hadoop is that it uses data locality to run MapReduce tasks on nodes close to where the file data resides.  In this fashion, networking bandwidth requirements are reduced and performance approximates local data access.

MapReduce programming is somewhat new, unique, and complex and was an outgrowth of Google’s MapReduce process.  As such, there have been a number of other Apache open source projects that have sprouted up namely, Cassandra, Chukya, Hbase, HiveMahout, and Pig to name just a few that provide easier ways to automatically generate MapReduce programs.  I will try to provide more information on these and other popular projects in a follow on post.

Hadoop fault tolerance

When an HDFS node fails, the NameNode detects a missing heartbeat signal and once detected, the NameNode will recreate all the missing replicas that resided on the failed DataNode.

Similarly, the MapReduce JobTracker node can detect that a TaskTracker has failed and allocate this work to another node in the cluster.  In this fashion, work will complete even in the face of node failures.

Hadoop distributions, support and who uses them

Alas, as in any open source project having a distribution that can be trusted and supported can take much of the risk out of using them and Hadoop is no exception. Probably the most popular distribution comes from Cloudera which contains all the above named projects and more and provides support.  Recently EMC announced that they will supply their own distribution and support of Hadoop as well. Amazon and other cloud computing providers also support Hadoop on their clusters but use other distributions (mostly Cloudera).

As to who uses Hadoop, it seems just about everyone on the web today is a Hadoop user, from Amazon to Yahoo including EBay Facebook, Google and Twitter to highlight just a few popular ones. There is a list on the Apache’s Hadoop website which provides more detail if interested.  The list indicates some of the Hadoop configurations and shows anywhere from a 18 node cluster to over 4500 nodes with multiple PBs of data storage.  Most of the big players are also active participants in the various open source projects around Hadoop and much of the code came from these organizations.

—-

I have been listening to the buzz on Hadoop for the last month and finally decided it was time I understood what it was.  This is my first attempt – hopefully, more to follow.

Comments?

EMCWorld day 3 …

Sometime this week EMC announced a new generation of Isilon NearLine storage which now includes HGST 3TB SATA disk drives.  With the new capacity the multi-node (144) Isilon cluster using the 108NL nodes can support 15PB of file data in a single file system.

Some of the booths along the walk to the solutions pavilion highlight EMC innovation winners. Two that caught my interest included:

  • Constellation computing – not quite sure how to define this but it’s distributed computing along with distributed data creation.  The intent is to move the data processing to the source of the data creation and keep the data there.  This might be very useful for applications that have many data sources and where data processing capabilities can be moved out to the nodes where the data was created. Seems highly scaleable but may depend on the ability to carve up the processing to work on the local data. I can see where compression, encryption, indexing and some statistical summarization can be done at the data creation site before it’s sent elsewhere. Sort of like both a sensor mesh with a processing nodes attached to the sensors configured as a sensor-proccessing grid.  Only one thing concerned me, there didn’t seem to be any central repository or control to this computing environment.  Probably what they intended, as the distributed solution is more adaptable and more scaleable than a centrally controlled environment.
  • Developing world healthcare cloud – seemed to be all about delivering healthcare to the bottom of the pyramid.  They won EMC’s social innovation award and are working with a group in Rwanda to try to provide better healthcare to remote villages.  It’s built around OpenMRS as a backend medical record archive hosted on EMC DC powered Iomega NAS storage and uses Google’s OpenDataKit to work with the data on mobile and laptop devices.  They showed a mobile phone which could be used to create, record and retrieve healthcare information (OpenMRS records) remotely and upload it sometime later when in range of a cell tower.  The solution also supports the download of a portion of the medical center’s health record database (e.g., a “cohort” slice, think a village’s healthcare records) onto a laptop, usable offline by a healthcare provider to update and record  patient health changes onsite and remotely.  Pulling all the technology together and delivering this as an application stack usable on mobile and laptop devices with minimal IT sophistication, storage and remote/mobile access are where the challenges lie.

Went to Sanjay’s (EMC’s CIO) keynote on EMC IT’s journey to IT-as-a-Service. As you can imagine it makes extensive use of VMware’s vSphere, vCloud, and vShield capabilities primarily in a private cloud infrastructure but they seem agnostic to a build-it or buy-it approach. EMC is about 75% virtualized today, and are starting to see significant and tangible OpEx and energy savings. They designed their North Carolina data center around the vCloud architecture and now are offering business users self service portals to provision VMs and business services…

Only caught the first section of BJ’s (President of BRS) keynote but he said recent analyst data (think IDC?) said that EMC was the overall leader (>64% market share) in purpose built backup appliances (Data Domain, Disk Library, Avamar data stores, etc.).  Too bad I had to step out but he looked like he was on a roll.

Comments?

One iPad per Child (OipC)

OLPC XO Beta1 (from wikipedia.org)
OLPC XO Beta1 (from wikipedia.org)

Starting thinking today that the iPad with some modifications  could be used to provide universal computing and information services to the world’s poor as a One iPad per Child (OipC).  Such a solution could easily replace the One Laptop Per Child (OLPC) that exists today with a more commercially viable product.

From my perspective only a few additions would make the current iPad ideal for universal OipC use.  Specifically, I would suggest we add

  • Solar battery charger – perhaps the back could be replaced with a solar panel to charge the battery.  Or maybe the front could be reconfigured to incorporate a solar charger underneath or within its touch panel screen.
  • Mesh WiFi – rather than being a standard WiFi target, it would be more useful for the OipC to support a mesh based WiFi system.  Such a mesh WiFi could route internet request packets/data from one OipC to another, until a base station were encountered providing a broadband portal for the mesh.
  • Open source free applications – it would be nice if more open office applications were ported to the new OipC so that free office tools could be used to create content.
  • External storage  – software support for NFS or CIFS over WiFi would allow for a more sophisticated computing environment and together with the mesh WiFi would allow a central storage repository for all activities.
  • Camera – for photos and video interaction/feedback.

    iPad (from wikipedia.org)
    iPad (from wikipedia.org)

Probably other changes needed but these will suffice for discussion purposes. With such a device and reasonable access to broadband, the world’s poor could easily have most of the information and computing capabilities of the richest nations.  They would have access to the Internet and as such could participate in remote k-12 education as well as obtain free courseware from university internet sites.  They would have access to online news, internet calling/instant messaging and free email services which could connect them to the rest of the world.

I believe most of the OipC hardware changes could be viable additions to the current iPad with the possible exception of the mesh WiFi.  But there might be a way to make a mesh WiFi that is software configurable with only modest hardware changes (using software radio transcievers).

Using the current iPad

Of course, the present iPad without change could be used to support all this, if one were to add some external hardware/software:

  • An external solar panel charging system – multiple solar charging stations for car batteries exist today which are used in remote areas.  If one were to wire up a cigarette lighter and purchase a car charger for the iPad this would suffice as a charging station. Perhaps such a system could be centralized in remote areas and people could pay a small fee to charge their iPads.
  • A remote WiFi hot spot – many ways to supply WiFi hot spots for rural areas.  I heard at one time Ireland was providing broadband to rural areas by using local pubs as hot spots.  Perhaps a local market could be wired/radio-connected to support village WiFi.
  • A camera – buy a cheap digital camera and the iPad camera connection kit.  This lacks real time video streaming but it could provide just about everything else.
  • Apps and storag – software apps could be produced by anyone.  Converting open office to work on an iPad doesn’t appear that daunting except for the desire to do it.  Providing external iPad storage can be provided today via cloud storage applications.  Supplying pure NFS or CIFS support as native iPad facilities that other apps could use would be more difficult but could be easily provided if there were a market.

The nice thing about the iPad is that it’s a monolithic, complete unit. Other than power there are minimal buttons/moving parts or external components present.  Such simplified componentry should make it more easily usable in all sorts of environments.  Not sure how rugged the current iPad is and how well it would work out in rural areas without shelter, but this could easily be gauged and changes made to improve it’s surviveability.

OipC costs

Having the mesh, solar charger, and camera all onboard the OipC would make this all easier to deploy but certainly not cheaper.  The current 16GB iPad parts and labor come in around US$260 (from livescience).  The additional parts to support the onboard camera, WiFi mesh and solar charger would drive costs up but perhaps not significantly.  For example, adding the iPhone 3m pixel camera to the iPad might cost about US$10 and a 3gS transciever (WiFi mesh substitute) would cost an additional US$3 (both from theappleblog).

As for the solar panel battery charger, I have no idea, but a 10W standalone solar panel can be had from Amazon for $80.  Granted it doesn’t include all the parts needed to convert power to something that the iPad can use and it’s big, 10″ by 17″.  This is not optimal and would need to be cut in half (both physically and costwise) to better fit the OicP back or front panel.

Such a device might be a worthy successor to OLPC at the cost of roughly double that devices price of US$150 per laptop.  Packaging all these capabilities in the OicP might bring some economies of scale that could potentially bring its price down some more.

Can the OipC replace the OLPC?

One obvious advantage that the OipC would have over the OLPC is that it was based on a commercial device.  If one were to use the iPad as it exists today with the external hardware discussed above it would be a purely commercial device.  As such, future applications should be more forthcoming, hardware advances should be automatically incorporated in the latest products, and a commercial market would exist to supply and support the products.  All this should result in better, more current software and hardware technology being deployed to 3rd world users.

Some disadvantages for the OipC vs. the OLPC include lack of a physical keyboard, open source operating system and access to all open source software, and usb ports.  Of course all the software and courseware specifically designed for the OLPC would also not work on the OipC.  The open sourced O/S and the USB are probably the most serious omissions. iPad has a number of external keyboard options which can be purchased if needed.

Now as to how to supply broadband to rural hot spots around the 3rd world, we must leave this for a future post…

Free P2P-Cloud Storage and Computing Services?

FFT_graph from Seti@home
FFT_graph from Seti@home

What would happen if somebody came up with a peer-to-peer cloud (P2P-Cloud) storage or computing service.  I see this as

  • Operating a little like Napster/Gnutella where many people come together and share out their storage/computing resources.
  • It could operate in a centralized or decentralized fashion
  • It  would allow access to data/computing resources anywhere from the internet

Everyone joining the P2P-cloud would need to set aside computing and/or storage resources they were willing to devote to the cloud.  By doing so, they would gain access to an equivalent amount (minus overhead) of other nodes computing and storage resources to use as they see fit.

P2P-Cloud Storage

For cloud storage the P2P-Cloud would create a common cloud data repository spread across all nodes in the network:

  • Data would be distributed across the network in such a way that would allow reconstruction within any reasonable time frame and would handle any reasonable amount of node outages without loss of data.
  • Data would be encrypted before being sent to the cloud rendering the data unreadable without the key.
  • Data would NOT necessarily be shared, but would be hosted on other users systems.

As such, if I were to offer up 100GB of storage to the P2P-Cloud, I would get at least a 100GB (less overhead) of protected storage elsewhere on the cloud to use as I see fit.  Some % of this would be lost to administration say 1-3% and redundancy protection say ~25% but the remaining 72GB of off-site storage could be very useful for DR purposes.

P2P-Cloud storage would provide a reliable, secure, distributed file repository that could be easily accessible from any internet location.  At a minimum, the service would be free and equivalent to what someone supplies (less overhead) to the P2P-Cloud Storage service.  If storage needs exceeded your commitment, more cloud storage could be provided at a modest cost to the consumer.  Such fees would be shared by all the participants offering excess [=offered – (consumed + overhead)] storage to the cloud .

P2P-Cloud Computing

Cloud computing is definitely more complex, but generally follows the Seti@HOME/BOINC model:

  • P2P-Cloud computing suppliers would agree to use something like a “new screensaver” which would perform computation while generating a viable screensaver.
  • Whenever the screensaver was invoked, it would start execution on the last assigned processing unit.  Intermediate work results would need to be saved and when completed, the answer could be sent to the requester and a new processing unit assigned.
  • Processing units would be assigned by the P2P-Cloud computing consumer, would be timeout-able and re-assignable at will.

Computing users won’t gain much if the computing time they consume is <= the computing time they offer (less overhead).  However, computing time offset may be worth something, i.e., computing time now might be more valuable than computing time tonite.  Which may offer a slight margin of value to help get this off the ground.  As such, P2P-Cloud computing suppliers would need to be able to specify when computing resources might be mostly available along with the type, quality and quantity.

Unclear how to secure the processing unit and this makes legal issues more prevalent.  That may not be much of a problem, as a complex distributed computing task makes little sense in isolation. But the (il-)legality of some data processing activities could conceivably put the provider in a precarious position. (Somebody from the legal profession would need clarify all this, but I would think that some “Amazon C2” like licensing might offer safe harbor here).

P2P-Cloud computing services wouldn’t necessarily be amenable to the more normal, non-distributed or linear computing tasks but one could view these as just a primitive version of distributed computing tasks.  In either case, any data needed for computation would need to be sent along with the computing software to be run on a distributed node.  Whether it’s worth the effort is something for the users to debate.

BOINC can provide a useful model here.  Also, the Condor(R) project at U. of Wisconsin/Madison can provide a similar framework for scheduling the work of a “less distributed” computing task model.  In my mind, both types of services ultimately need to be provided.

To generate more compute servers, the SETI@Home and similar BOINC projects rely on doing good deeds.  As such, if you can make your computing task  do something of value to most users then maybe that’s enough. In that case, I would suggest joining up as a BOINC project. For the rest of us, doing more mundane data processing, just offering our compute services to the P2P-Cloud will need to suffice.

Starting up the P2P-Cloud

Bootstrapping the P2P-Cloud might take some effort but once going it should be self sustaining (assuming no centralized infrastructure).  I envision an open source solution, taking off from the work done on Napster&Gnutella and/or Boinc&Condor.

I believe the P2P-Cloud Storage service would be the easiest to get started.  BOINC and SETI@home (list of active Boinc projects) have been around a lot longer than cloud storage but their existence suggests that with the right incentives, even the P2P-Cloud Computing service can make sense.