Microsoft has added a Basic tier to Azure Search, bringing the total number of available plans to four. Launched nearly a year ago and influenced by the software giant's Bing public search engine, the "search-as-a-service" offering enables Web and mobile developers to build custom, natural-language search experiences for their applications.
Costing $75 per month—half off during the preview period—a Basic Azure Search plan can index up to 1 million documents. For comparison's sake, the Free plan tops out at 10,000 documents and 50 megabytes of storage, while Standard S1 subscriptions (approximately $250 per month) top out at 180 million documents and 300GB of storage. A Standard S2 plan is also available, offering even more capacity.
The new Basic tier is Microsoft's response to customer demand for a more modest alternative to the Standard plans, said Liam Cavanagh, principal program manager of Microsoft Azure Search, in a March 2 announcement. "Basic is great for cases where you need the production-class characteristics of Standard but have lower capacity requirements," he stated.
Those production-class capabilities include dedicated partitions and service workloads (replicas), along with resource isolation and service-level agreement (SLA) guarantees, which are not offered in the Free tier. Hence, Microsoft discourages the use of Free plans for workloads meant to power Web and mobile app search experiences in production.
In terms of performance, Microsoft was able to index roughly 15,000 documents per minute when organized into batches of 1,000 documents each, revealed Cavanagh. Queries will slow down if pushed to those extremes, he cautioned, but customers should be able to index a million documents in just over an hour. In its tests, Microsoft used documents of around 1KB in size and a nine-field index with a mix of facetable (used in Azure Search's faceted navigation feature), filterable and searchable fields, he noted.
Cavanagh's group was able to complete five queries per second with latencies of approximately 200 milliseconds for searches that returned hundreds of documents. Searches that generate tens of thousands of results will experience a drop in query-per-second performance and latencies that hover around the 300- to 400-millisecond range. To maintain brisk performance with a Basic Azure Search plan, Cavanagh recommends reusing HTTP clients to avoid extra latency caused by re-establishing connections and avoiding queries that result in a large number of document matches.
Targeting a greater diversity of workloads and clinging tight to its cloud-first approach to enterprise software and services, Microsoft has steadily been expanding its cloud's search capabilities.
Last month, the company turned its attention to unstructured data with a beta version of its search indexer for Azure Blob Storage. The indexer can extract text and metadata from Office file formats (DOCX/DOC, XLSX/XLS, PPTX/PPT and MSG) along with EML, HTML, XML, ZIP and plain text files.