When it comes to search, it can never be fast enough

SRCH2 in-memory database search provides compelling speed advantages when compared to Solr.

SRCH2 recently announced the results of some in-house benchmark testing, and they are impressive.  In comparing their in-memory search (SRCH2 4.3.2) solution to SOLR (Apache Solr 4.8.0), the results pointed to SRCH2 performance throughput increase.  The details of the report can be found here on their website. Highlights of the report are as follows:

• SRCH2 performed 5.4x faster than a Solr system with 5m indexed records.
• SRCH2 had 4x query throughput in the test.
• SRCH2 performance improved as more documents were indexed.

As seen on their graph (below) summarizing the results: SRCH2 vs. Solr Benchmark Reports 4x Throughput Increase.

What does this mean in practical application?

There is a hidden relevancy difference because SRCH2 provides additional index structures and encoding innovations that are not available in SOLR. These include an encoded two-way index, incremental caching, and geo-location awareness.  This means SRCH2 results would be faster and more relevant when compared to SOLR results generated from a comparable corpus. This has a significant advantage as you scale out and scale down. There two practical applications:

• Autocomplete/suggested search
• Mobile search

Google Instant Search is an autocomplete service which many of our customers aspire to and starting with SRCH2 can aid them in that.  Separately SRCH2 has released an Android SDK and has worked hard to get a search system to perform in smaller environments where memory, power, and storage are constrained. We think SRCH2 is worth checking out.  You can download the full report here.