change img syntax to markdown instead of html
diff --git a/src/pages/technology.md b/src/pages/technology.md
index 0a9d6ed..b09d840 100644
--- a/src/pages/technology.md
+++ b/src/pages/technology.md
@@ -11,9 +11,7 @@
Apache Druid is used to power real-time analytics applications that require fast queries at scale and under load on streaming and batch data. Druid features a unique distributed architecture across its ingestion, storage, and query layer to handle the scale needed for large aggregations with the performance needed for applications.
-<div class="image-large">
- <img src="img/diagram-7.png" />
-</div>
+
Druid is a services-based architecture that consists of independently scalable services for ingestion, querying, and orchestration, each of which can be fine-tuned to optimize cluster resources for mixed use cases and workloads. For example, more resources can be directed to Druid’s query service while providing less resources to ingestion as workloads change. Druid services can fail without impact on the operations of other services.
@@ -28,7 +26,7 @@
In addition, Druid utilizes a deep storage layer - cloud object storage or HDFS - that contains an additional copy of each data segment. It enables background data movement between Druid processes and also provides a highly durable data source to recover from system failures.
-For more information, please visit [our docs page](/docs/latest/design).
+For more information, please visit [our docs page](../../docs/latest/design).
## Ingestion Layer
@@ -37,9 +35,7 @@
When data is ingested into Druid, it is automatically indexed, partitioned, and, optionally, partially pre-aggregated (known as <a href="https://druid.apache.org/docs/latest/tutorials/tutorial-rollup.html">"rollup"</a>). Compressed bitmap indexes enable fast filtering and searching across multiple columns. Data is partitioned by time and, optionally, by other dimensions.
-<div class="image-large">
- <img alt="Stream Ingestion Layer" src="img/ingestion_layer_stream_batch.png" />
-</div>
+
### Stream Data
@@ -49,7 +45,7 @@
Druid easily ingests data from object stores including HDFS, Amazon S3, Azure Blob, and Google Cloud Storage as well as data files from databases and other sources. The data files can be in a number of common formats, including JSON, CSV, TSV, Parquet, ORC, Avro, and Protobuf. Druid supports both SQL batch import and in-database transformations.
-For more information, please visit [our docs page](/docs/latest/ingestion/index.html).
+For more information, please visit [our docs page](/docs/latest/ingestion/).
## Storage Format
@@ -57,11 +53,9 @@
Within the segments, data storage is column-oriented. Queries only load the specific columns needed for each request. Each column’s storage is optimized by data type, which further improves the performance of scans and aggregations. String columns are stored using compressed dictionary encoding, while numeric columns are stored using compressed raw values.
-<div class="image-large">
- <img alt="Graphical User Interface, Application" src="img/graphical_ui_application_v2.png" />
-</div>
+
-For more information, please visit [our docs page](/docs/latest/design/segments).
+For more information, please visit [our docs page](../../docs/latest/design/segments).
## Interactive Queries
@@ -69,12 +63,10 @@
With this engine, Druid only reads from segments that are pre-loaded into memory or local storage in the data nodes. This ensures fast performance as data is co-located with computing resources and does not have to move across a network. Data is then queried using scatter/gather for optimal parallelization.
-<div class="image-large">
- <img alt="Interactive Querying Scatter Gather Diagram" src="img/scatter_gather_diagram.png" />
-</div>
+
First, the query engine prunes the list of segments, creating a list of which segments are relevant to the query based on time-internals and other filters. Next, queries are divided into discrete pieces and sent in parallel to the data nodes that are managing each relevant segment or copy of that segment (“scatter”). On the data nodes, the sub-queries are processed and sent back to the query nodes to merge the final result set (“gather”).
Scatter/gather works from the smallest single server cluster (all of Druid on one server) to clusters with thousands of servers, enabling sub-second performance for most queries even with very large data sets of multiple petabytes.
-For more information, please visit [our docs page](/docs/latest/querying/).
\ No newline at end of file
+For more information, please visit [our docs page](../../docs/latest/querying/).
\ No newline at end of file
