EXASolution is flexible enough to be suited to a number of user environments that vary from departmental installations up to corporate-level implementations, and ranging from conventional business intelligence environments to real-time analytics. However, EXASOL is currently focusing on environments where there is a need is to run complex, analytic queries in a timely manner. In these sorts of environments EXASolution has a number of potential advantages over its rivals, which range from ease of implementation to performance and from low cost of ownership to rapid time to value.
EXASOL is fairly agnostic with respect to industry verticals although it has enjoyed success in the online gaming and new media sector, manufacturing, telecommunications, market research, financial services and retail sectors.
The most notable of its publicly named customers are Xing, Adidas, Groupon, Olympus, The Co-op, Webtrekk and Sony Music.
EXASolution from EXASOL is a massively parallel database management system especially developed for data warehousing applications, which is available as a hosted service (EXACloud), as a software only solution or as a pre-configured appliance running on commodity Intel-based hardware. EXASolution runs on its own operating system, EXACluster OS, to underpin EXASolution, with a single install for the entire environment. EXACluster OS is an application layer that sits on top of a standard Linux distribution and offers a number of advantages when compared to conventional approaches to clustering (on which EXASolution is based) particularly with respect to cluster management and installation.
As of the time of writing, EXASolution 4.0, running on Dell hardware, holds the performance records for the TPC-H benchmark at every level up to 10Tb as well as having the best price/performance records at each of these levels. This doesn't mean that EXASolution isn't capable of supporting more (much more) disk than this but that TPC-H benchmarking is only competitive up to this level (there is only one published figure that is larger than this and that was published in 2007). The one caveat we would make to this otherwise impressive set of results is that benchmarks do not necessarily equate to the workload of any particular company. Nevertheless, EXASolution can provide answers to queries (particularly complex queries) much faster than by means of conventional data warehousing. There are a number of technical reasons for this, including the use of column-based storage, in-memory processing and automated index creation.
In practice, the scalability of EXASolution is dependent on memory rather than disk capacity because the product makes extensive use of in-memory processing. The company has successfully tested warehouses with up to 5Tb of memory, which would equate to anything between 50 and 200Tb of raw data, depending on usage. The other important aspect of scalability is the number of users that can be supported and the number of concurrent queries. Existing EXASOL customers have as many as 3,000 of the former running up to 150 concurrent queries at peak periods. Another important aspect of the product is its self-tuning concept. Customers don't need to bother with schema design constraints, indices or other optimising processes. The Engine takes care of all these tasks automatically.
An additional product, EXAPowerlytics, is available and offers customers the ability to implement their own parallel algorithms within the EXASolution cluster, natively integrated into normal SQL processing. This framework also includes "R" as a programming language, Map/Reduce capabilities and the processing of unstructured data which can be dynamically loaded from external sources like through the Facebook API or Hadoop clusters. We understand that further integration with Hadoop will be a major feature of the next product release.
EXASOL provides the sort of services one would expect: consultancy, support, training and so forth. Service partners such as SSH Viveon, Logica and The Intelligent Edge Group provide additional support.