Raima Product Releases & Patches

RDM Embedded
Version	Release Date
9.1 - J	2/9/2011
7.2 - J	3/16/2011
10.0.5	3/28/2011
10.0.6	5/10/2011

RDM Server
Version	Release Date
8.0 - R	2/4/2011
8.3 - L	2/25/2011
8.1.28	3/8/2011
6.0 - Y	3/29/2011
8.0 - S	3/31/2011
8.1.29	4/4/2011
8.3 - M	4/8/2011
8.3 - N	5/20/2011
8.1.30	4/29/2011

Visit Our New Website! We have a fresh new look and hope that you find it easy to navigate and quickly find what you need. New material and updates will be continually added. For the latest in what’s happening at Raima, visit www.raima.com.

In this Edition

· RDM Embedded 10.1 Release Candidate

· Raima on the App Store

· Test Bench: SQLite & RDMe 10.0

· Test Bench: PostgreSQL & RDMe 10.0

· Recent Releases & Patches

· New White Papers

· Training Calendar

Raima putting SQLite and RDMe 10.0 on the Test Bench

Raima now has an app published to the Apple AppStore! As a project to demonstrate the
performance of RDM Embedded on an iOS device, we created an app called AirportSearch.
This app allows you to search from a database of over 5,000 airports, 9,000 airlines, and
over 56,000 routes between them instantly. Available now for free, check it out at
http://itunes.apple.com/us/app/airportsearch-free/id436851243?mt=8. In internal tests, RDMe has been shown to have excellent performance on iOS devices, as well as bringing powerful features such as DataFlow and Distributed databases to a mobile device. Watch for official support in the RDM Embedded SDK for the iOS platform later this year in the
upcoming release of RDMe 10.1!

Raima on the App Store

Free Software, SQLite, Not Free to Grow with Your System

We admit it -- SQLite is easy to obtain and start using. It is
 "good enough" for your needs, so why not go with it? In
many cases, you may be right. But if your database is likely
to be used by 4 or more people at once, you may find
yourself locked into a situation that degenerates as it
grows. As the graph shows, the SQLite transaction
throughput starts out really good -- even better than RDM
Embedded version 10, for 2, 3 and 4 users (and that’s
saying something!). But add a couple more users and the
story starts to change. In database terms, this is a nosedive,
as SQLite goes from over 8000 transactions per minute
system-wide, down to less than 600 per minute. Another
way to look at this is the time each user waits for a
transaction to complete. The best is 16 milliseconds (good!) when there are only two users, but it degenerates to nearly 3 seconds (awful) when there are 24 users.

To receive a full report showing how RDM Embedded 10 and SQLite respond to various loads, and how well they scale up on a multi-core, multi-disk-drive computer, email sales@raima.com. Don’t limit your software. For more information about RDM Embedded visit our website.

New White Papers:

Data Management Solution: Build vs. Buy

This article explores "Build vs. Buy", a key decision for determining the best strategy for meeting your product’s embedded data
management needs. Get It Now!

 

Embedded Databases in the Mobile Environment

This white paper addresses a variety of markets and the growing use of embedded database management in the mobile
environment. Get It Now!

Training

September 6th-8th, 2011 RDM Embedded Training in Seattle

RDM Embedded Jump Start Training is designed to help software
engineers get up and running with the RDM Embedded database management software. Training is tailored to the attendee’s database expertise. Read More

Contact Your Raima Sales Team

For more information about Raima products and services and finding your embedded database management solution:

Email: sales@raima.com  
Web:  Contact Us 
Phone: +1 206 748 5300 (Worldwide), +44 1628 826 800 (Europe)

RDM Embedded 10.1 Coming Soon! Register to be part of the Release Candidate (RC) Program

Raima putting PostgreSQL and RDMe 10.0 on the Test Bench

As we put the finishing touches on RDM Embedded 10.1 we are looking for those of you who would like to receive a copy of the release candidate in the next few weeks. This release of RDM Embedded includes a brand new SQL API designed to meet the needs of today’s embedded systems. A C++ interface and hash indices, which in tests of large databases is showing triple digit percentage increases in performance. If you would like to receive a copy of this release candidate, and have the time to try it out, please register at www.raima.com/RCProgram .

Off the WAL Performance

Write-Ahead-Logging (WAL) is a technique used by
PostgreSQL and SQLite to allow database reading to occur
in parallel with transactions, safely without locks. WAL is used
to implement multi-version-concurrency-control (MVCC). We
have observed that the read times are much improved when
this method is used, because locking is always expensive.

RDM Embedded 10 also implements MVCC, but not through
WAL. There's not really a name for the Raima MVCC
technique, so we’ll just call it the "live cache" method for the
sake of this article. Our Transactional File Server (TFS)
component caches before and after images of database
pages when a process is reading with MVCC, so the
"snapshot" required by the reader is instantly available.  We didn’t realize how effective this technique was until we started comparing it to PostgreSQL and SQLite.

See the graph for a representative result of RDM Embedded to PostgreSQL.

While reading is quite quick when there is no load of transactions, the WAL technique slows down significantly as the
transaction load increases. Not so with the "live cache" method. Our benchmarking so far shows almost no slowdown for readers when the database is being actively updated. The PostgreSQL reading ranges from .91 milliseconds to 120
milliseconds.  Doing the same work, RDM Embedded 10 reading ranges from .69 milliseconds to .69 milliseconds. That's no change, for those of you doing the math. For more information about RDM Embedded visit our website.