0

help me understand what are the advantages of hadoop over teradata. Why should we migrate from teradat to hadoop. In my applications I have some reports retrieving data from teradata, reports are very slow because of millions of row data.

Will migrating to hadoop resolve it?

0m3r
  • 12,286
  • 15
  • 35
  • 71
aru
  • 59
  • 1
  • 7
  • 16
  • How complex are your reports, how many joins, aggregates etc.? – dnoeth Aug 03 '15 at 14:49
  • reports are quite complex.currently report is fetching data from teradata view . this view is created on two teradata tables having hundreads of columns and millions of rows. – aru Aug 12 '15 at 08:24
  • *Millions of rows* is not much (the biggest Teradata table at Ebay holds 2.600.000.000.000 rows), unless that view is very complex it shouldn't run that long. So before migrating you might try to tune this view/report first, you might open a new question and post some more details :-) – dnoeth Aug 12 '15 at 08:46

1 Answers1

2

Possible duplicate of hadoop vs teradata what is the difference.

The main advantage of Hadoop system is scalability with commodity hardware.

As pointed out by @dnoeth in comments. Teradata also scales out similar to Hadoop. But it can only scale out using expensive servers. However Hadoop systems can scale out using any commodity hardware (more commonly available less expensive hardware).

In pioneer days they used oxen for heavy pulling, and when one ox couldn’t budge a log, they didn’t try to grow a larger ox. We shouldn’t be trying for bigger computers, but for more systems of computers.
—Grace Hopper

Hadoop Advantages

  1. Fault tolerance provided as part of the system. Graceful degradation, and data availability taken care of.
  2. Individual nodes in the cluster can vary in their capacities.
  3. Flexibility to add/remove nodes from cluster without shutting the cluster down.

Hadoop Disadvantages

  1. It is batch processing system with high throughput and high latency.
  2. Hadoop distributed file system doesn't allow modifying existing files.
  3. Performance is very poor if used for small data.
Community
  • 1
  • 1
shanmuga
  • 4,329
  • 2
  • 21
  • 35
  • 3
    Teradata doesn't scale up (single server), it scales out (multiple servers), quite similar to Hadoop (although on more expensive servers, up to 2*14 core Xeon and 512 GB RAM). In fact the basic principles are the same for both, split your work in lots of smaller subsets and process them in parallel. Teradata is doing MapReduce since over 30 years :-) – dnoeth Aug 03 '15 at 14:34
  • @dnoeth thanks for explaining. I'll change my answer accordingly. – shanmuga Aug 04 '15 at 05:47