Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Perimeter

12/7/2012
01:44 PM
Adrian Lane
Adrian Lane
Commentary
50%
50%

What Is Big Data?

Big data is not about buying more big iron

When someone says big data, what do you think of?

Do you think of mainframes? Data warehouses? Do you think of Oracle Grids, Exadata, or Teradata clusters?

Perhaps you think of Hadoop, MongoDB, Cassandra, or CouchDB? Or maybe it's any NoSQL database?

Or perhaps you think it's just a giant mass of data in one place?

If you read press articles on big data, then it's all of these things. It's my belief that no good definition of big data exists today. In fact, the term is so overused, and I think intentionally so, that it's almost meaningless. I want to address that problem here.

And I'll state up front that the big data phenomena is not because people are buying more big iron.

During the past year, I've spent an inordinate amount of time researching security in and around big data clusters. It has been a challenge; each time I think I have a handle on one aspect of what constitutes big data, I find an exception that breaks the conceptual model I've created. Every time I think I've quantified a specific attribute or feature, I find another variation of NoSQL that's an exception to the rule. It was even a struggle to just define what big data actually is, with definitions from Wikipedia and other sources missing several essential ingredients: In fact, the definition section of the Wikipedia entry on big data does not really offer a definition at all. All in all, this is one of the most difficult, and interesting, research projects I've been involved with.

I want to share some of the results of that research here because I think it will be helpful in understanding why securing big data is difficult, and how the challenge is not the same as relational platforms many of you are familiar with. In a future post, I'll discuss some of the fundamental differences in how big data systems are deployed and managed from a security perspective, but before I can talk about how to secure "it," I need to define what "it" is.

Yes, big data is about lots of data, of differing types, coming in at velocities that cripple most traditional database systems. But there are other essential characteristics besides size and the need for fast insertion, such as the ability to elastically scale as the data set grows. It's about distributed, parallel processing to tackle massive analysis tasks. It's about data redundancy to provide failure resistant operation, which is critical when computing environments span so many systems that hardware failures are to be expected during the course of operation.

And just as importantly, these systems are hardware-agnostic, accessible from complexity standpoint, extensible, and relatively inexpensive. These characteristics define big data systems.

The poster child for big data is Hadoop, which is a framework that at its core provides data management and query (map-reduce) services across (potentially) thousands of servers. Everything about big data clusters is designed to address storage and processing of multiple terabytes of data across as many systems as needed, in an elastic, expansive way. In fact, these clusters are so large that the prospect or failure increases to the point where it's probable a node will fail. Without elasticity, resiliency, and potential to process requests in more than one location, that makes big data different than the databases that have come before it.

But the reason why big data is a major trend is because of the convergence of three things: huge amounts of data with cheap computing resources and free (or nearly free) analytic tools. Enterprises and midmarket firms are all embracing big data not because they can suddenly afford to invest millions of dollars in data warehouse systems, MPPs, mainframes, or giant systems in-a-box. It's because they can now afford data analysis on massive data sets without spending much money up front. Cheap, commodity, or cloud computing resources with free and easy data management systems like Hadoop make it possible.

If you need to understand what big data is, then consider the characteristics outlined above. They should help you differentiate traditional systems from big data.

Adrian Lane is an analyst/CTO with Securosis LLC, an independent security consulting practice. Special to Dark Reading. Adrian Lane is a Security Strategist and brings over 25 years of industry experience to the Securosis team, much of it at the executive level. Adrian specializes in database security, data security, and secure software development. With experience at Ingres, Oracle, and ... View Full Bio

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Cryptodd
50%
50%
Cryptodd,
User Rank: Moderator
12/10/2012 | 5:49:16 PM
re: What Is Big Data?
Adrian from Securosis did a fantastic piece of research on securing Big Data that provides a nice summary of the topic.- The paper is available for download at http://www.vormetric.com/resou...-. Enjoy, TT
COVID-19: Latest Security News & Commentary
Dark Reading Staff 9/25/2020
Hacking Yourself: Marie Moe and Pacemaker Security
Gary McGraw Ph.D., Co-founder Berryville Institute of Machine Learning,  9/21/2020
Startup Aims to Map and Track All the IT and Security Things
Kelly Jackson Higgins, Executive Editor at Dark Reading,  9/22/2020
Register for Dark Reading Newsletters
White Papers
Video
Cartoon
Current Issue
Special Report: Computing's New Normal
This special report examines how IT security organizations have adapted to the "new normal" of computing and what the long-term effects will be. Read it and get a unique set of perspectives on issues ranging from new threats & vulnerabilities as a result of remote working to how enterprise security strategy will be affected long term.
Flash Poll
How IT Security Organizations are Attacking the Cybersecurity Problem
How IT Security Organizations are Attacking the Cybersecurity Problem
The COVID-19 pandemic turned the world -- and enterprise computing -- on end. Here's a look at how cybersecurity teams are retrenching their defense strategies, rebuilding their teams, and selecting new technologies to stop the oncoming rise of online attacks.
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2020-15208
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, when determining the common dimension size of two tensors, TFLite uses a `DCHECK` which is no-op outside of debug compilation modes. Since the function always returns the dimension of the first tensor, malicious attackers can ...
CVE-2020-15209
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, a crafted TFLite model can force a node to have as input a tensor backed by a `nullptr` buffer. This can be achieved by changing a buffer index in the flatbuffer serialization to convert a read-only tensor to a read-write one....
CVE-2020-15210
PUBLISHED: 2020-09-25
In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, if a TFLite saved model uses the same tensor as both input and output of an operator, then, depending on the operator, we can observe a segmentation fault or just memory corruption. We have patched the issue in d58c96946b and ...
CVE-2020-15211
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, saved models in the flatbuffer format use a double indexing scheme: a model has a set of subgraphs, each subgraph has a set of operators and each operator has a set of input/output tensors. The flatbuffer format uses indices f...
CVE-2020-15212
PUBLISHED: 2020-09-25
In TensorFlow Lite before versions 2.2.1 and 2.3.1, models using segment sum can trigger writes outside of bounds of heap allocated buffers by inserting negative elements in the segment ids tensor. Users having access to `segment_ids_data` can alter `output_index` and then write to outside of `outpu...