Requirements Before Installing Sintelix

We recommend storing the Sintelix database on an SSD. This significantly increases the speed of network formation (which is computationally intensive), especially for large networks, and makes search functions at least an order of magnitude faster.

The Java Virtual Machine (JVM) created to host Sintelix is designed to limit the usage of machine RAM to not more than 75% less 512 MB. This means that a 4-core deployment would require at least 11.17 GB of RAM. A 6-core deployment would require at least 15.17 GB of RAM.

Multiprocessor boards enable effective numbers of cores up to 256.

Where the Sintelix program is installed.

Windows Default : c:\Program Files (x86)\Sintelix.

This location is where source documents are loaded into the Sintelix database.

Accessing documents from the Server Library is the fastest way to load a large number of documents but requires a common file system directory where:

• users can copy documents, and
• Sintelix has read access.

The folder where the Sintelix database is installed.

Windows Default : c:\Sintelix database.

Sintelix makes frequent writes to its database location. Certain anti-virus software will automatically scan all the files written by Sintelix, which slows down Sintelix. It is highly recommended you exclude the Sintelix database location from real-time anti-virus scans. No executable or dangerous files are written there.

Sintelix has both document and network formation workflows, which can be run in series to provide incremental network formation.

The time taken to process and store documents increases linearly with the number of documents stored.

The time taken to create a network of any size increases faster than linearly (approximately the number of entities included to the power of 1.4).

Sintelix has both document and network formation work flows, which can be run in series to provide incremental network formation. Approximate processing speeds for collections and networks are:

• Entity extraction (processing documents for storage and search in Sintelix)
• 30 pages of text per core per second.
• 2.6 million pages per core per day.
• 80 million entities per core per day.

• Entity extraction with entity resolution (creating entity networks/graphs)
• 10 pages of text per core per second.
• 860,000 pages per core per day.
• 26 million entities per core per day.
• roughly 50 GB of electronic (that is, not scanned) documents per core per day, but this varies greatly with document type (PDF tends to be fast, DOC are slow).
• network formation gets progressively slower as the number (n) of entities increases (~n^1.4)

• Scanned documents (requiring OCR Optical Character Recognition, a method of converting images of typed, printed or handwritten text into machine-readable text.)
• approximately 1 second per core per page, that is, very slow (varies with the OCR server used).

Sizing of storage requires care.

Sintelix document collections and networks are not limited in size by software considerations. All you need is enough disk space (see below).

Document processing deployments are where nothing is saved to a Sintelix database.

Stateless use of Sintelix is only relevant for document processing where each document is sent separately to Sintelix via the Document Processing web service. Sintelix returns the facts (entities, relationships, metadata) separately for each document. No information in the documents or calculated from the documents is stored in Sintelix.

Windows:

The Windows installation deploys with its own version of Java used only for Sintelix within the Sintelix server.

Linux:

It is recommended to upgrade the browser to the latest version as soon as upgraded versions are made available. Later versions of browsers often have higher levels of security and better application functionality.

Supported browsers include:

• Microsoft Edge
• Google Chrome
• Mozilla Firefox

Sintelix uses Google Cloud platform to access APIs.

Microsoft Azure is used by Media Processing Server for Transcription requirements.

If Sintelix is integrated into a third-party system, install Sintelix on the same machine as the third-party system or have a high-speed network between the computers hosting the two systems.

Sintelix is capable of fetching documents by URL, using the http or https protocol. If Sintelix is isolated from the relevant network (for example, the Internet) then this functionality will not work.

When you use Sintelix to browse documents with geographical information (location names, coordinates) the web browser may attempt to connect to the Internet to show detailed maps (depending on your map provider selection). High detail maps will fail to function without an Internet connection.

If there is an additional proxy server to isolate Sintelix from the network, please make sure it doesn’t impose severe upload size restrictions as uploading documents larger than the restriction will fail.

Depending on the Sintelix licensing agreement, Sintelix may need to establish a connection to the activation server. If this restriction is in place, Sintelix needs to be able to connect to the Internet to make the connection. Activation occurs on startup and every 24 hours. Note the instructions for configuring the firewall to allow access to the Sintelix server URL and port required to maintain activation. See Preparing to Install Sintelix.

Depending on the Sintelix licensing agreement, Sintelix may only start on computers with network adaptors that have a license-defined MAC address. If this restriction is in place, ensure that you install Sintelix on matching hardware. When providing a MAC address, we recommend using an Ethernet adapter address instead of a WiFi adapter address.

Sintelix utilises USB Dongles to support verification of Activation Keys when running Sintelix in an offline environment. Refer to Dongle Licensing.