Requirements

Iotellect Server can run under virtually any Java-enabled OS. It has been tested and verified to correctly run under the following operating systems:

• Windows 7, 10, 11, Server 2016, Server 2019 (32-bit and 64-bit versions)

• Linux (32-bit and 64-bit versions with Java support)

• ARM-based Linux

Other operating systems that are known to be compatible with Iotellect Server include AIX, Solaris, FreeBSD, etc.

Iotellect is a comprehensive device management solution capable of monitoring device networks of all sizes. Most Iotellect Server installations perform well on desktop-class systems. However, when managing larger networks or devices that generate a significant amount of data, you need to give additional consideration to the hardware and system configuration used.

Several factors can impact system scalability. The first is the number of devices that you will be monitoring with Iotellect. Performance tuning may be required when there are more than 100 devices. The second is the amount of data you collect from each device. The third factor to consider is monitoring time intervals. For example, if you set your synchronization time intervals to collect data every second instead of the default time intervals, Iotellect Server will have to work harder and system requirements will increase. Finally, the number of simultaneous operators accessing the system will directly impact performance.

When planning your Iotellect installation, keep in mind that CPU and memory usage is highly dependent on the types of managed devices and device drivers. In situations where many devices are controlled, it is recommended to use a dedicated, faster server.

Server CPU usage also depends on:

• Number of events (including variable value change events) routed via server core

• Number and complexity of server operations (e.g. alert trigger evaluations, model binding executions, or job scheduler runs)

• Number of concurrent user connections and complexity of user operations

Server RAM usage depends on:

• Number of connected devices

• Number of device settings, operations, and event types

• Number and complexity of data processing tools (alerts, reports, and, especially, widgets)

• Number of concurrent client connections

If you have additional questions or experience performance problems, please contact Iotellect to speak with one of our system engineers or visit Iotellect website.

When building a unique solution based on Iotellect, you should consider the platform like a "development environment" rather than a "box product". Thinking from this point brings you to a conclusion that real hardware requirements will fully depend on the logic of your product rather than formal parameters like device count or the number of incoming events per second. This unfortunately means that in such unique cases you might need a performance testing environment to evaluate hardware requirements for a production server. Example: A model with ten bindings executing once per second will cause near-zero CPU and memory impact if those bindings are just performing some calculations, e.g. evaluating values of custom device KPIs based on its raw metrics. However, the same bindings will cause high CPU load and memory usage if each of them is executing a query that fetches a million items from the unified model and performs some sorting/filtering/grouping of those items.

Planning disk space utilization should take into account that most of the disk space used by Iotellect Server is normally occupied by event storage. You can evaluate disk space usage by estimating the maximum number of events (including historical variable values) that will be stored persistently and multiplying it by an average event disk footprint that is around 1 Kb.

We provide several hardware configuration tables that are suitable for systems with low, medium, high, and very high per-device loads.

The below hardware requirements are just estimations. For example, a single controller providing ten thousand I/O channels that are polled at a 10 Hz rate can require a dedicated powerful Iotellect Server for adequate data processing and storage. On the contrary, one server can support 100 000 hardware Agents that connect to the server once per day to send several events each.

The below requirements for Disk Space assume that 100% of historical device data samples and events are stored within the local storage facility.

All memory requirements assume that the operating system occupies:

• 50% of RAM if server RAM is below 2 Gb

• 1 Gb of RAM if server RAM is above 2 Gb

Very low load implies an average of up to 30 settings/operations/events types per device, with a relatively small number of simple data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 100 event instances and metric updates per day. Disk space requirements assume that all historical values and events are stored for 1 year.

Low-level load implies an average of 30-100 settings/operations/events types per device, and an average number of data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 1000 event instances and metric updates per day. Disk space requirements assume that all historical values and events are stored for 1 year.

Medium-level load implies an average of 100-300 settings/operations/events types per device, with a large number of complex data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 10000 event instances and metric updates per day. Disk space requirements assume that all historical values and events are stored for 1 year.

High-level load implies an average of 300-1000 settings/operations/events types per device, with a large number of complex data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 100 000 event instances and metric updates per day. Typical devices of this type are controllers/gateways aggregating data from multiple field-level devices or sensors. Disk space requirements assume that all historical values and events are stored for 1 year.

A very high load implies an average of 1-3 thousand settings/operations/events types per device, with a large number of complex data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 10 000 000 event instances and metric updates per day. Typical devices of this type are other enterprise applications integrated with Iotellect as data sources and "exposing" multiple devices/data/events metrics managed by them through a single Iotellect device account. Disk space requirements assume that all historical values and events are stored for 1 year.

Extremely high load implies an average of over 3 thousand settings/operations/events types per device, with a large number of complex data processing tools (alerts, reports, trackers, models, widgets). Each device generates an average of less than 1 000 000 000 event instances and metric updates per day. Typical devices of this type are high-performance gateways and enterprise applications integrated with Iotellect as data sources and generating a high volume of data. Disk space requirements assume that all historical values and events are stored for 1 month.

If Iotellect Server is using an external database, hardware requirements for the database server machine are normally same as requirements for the machine running Iotellect Server itself. Thus, it means that database server hardware requirements depend to the number of devices connected to the Iotellect Server and an average per-device load.

Iotellect Server should be connected to the database server via a fast dedicated network link. Dedicated Gigabit Ethernet link with latency (ping time) of less than 5 milliseconds will ensure sufficient performance in almost every case.

If an external NoSQL database is used, Iotellect machine should be equipped with a 50 Gb solid-state drive (SSD) used for configuration storage. Disk space requirements should be applied to the external database server in this case.

Recommended memory size for an external NoSQL database server is 12-16 Gb.

Iotellect Server requires the following software to work:

JVM is the only absolute requirement to run the Iotellect Server. The JVM may be bundled with the installer itself or pre-installed on the target system (see installation). The minimal version of JVM required to run the Iotellect Server is Java 17.

Iotellect Server requires a database to operate. Distribution packages supplied by Iotellect include a simple relational database engine that is used by default. For large-scale systems we recommend using an industrial-grade stand-alone relational database or integrated NoSQL database. Iotellect Server is compatible with a wide range of third-party relational database servers:

• MySQL

• PostgreSQL

• Oracle

• Microsoft SQL Server

• Firebird

• and many others…

Iotellect Server and the database server don't have to be installed on the same server. It is also possible to configure Iotellect Server to work with a remote database, but in this case, its performance will depend on network conditions.

To use the Dynamic DNS Service you need an RFC-2136 compliant DNS Server. This can be a local or remote server. RFC-2136 (Request For Comments) specification is called "Dynamic Updates in the Domain Name System". Dynamic updates are supported by most present-day DNS Servers, including BIND and Windows 2000/2003 DNS service.

A mail server is required to allow Iotellect Server to send alert notification e-mails. Iotellect Server can communicate with virtually any SMTP mail server, including ones that require authentication.

Iotellect Server listens various port numbers in order to accept incoming TCP connections and UDP data from various devices, data sources and systems.

Was this page helpful?

Tell us more

Send

Thank you!