How to Select a Hardware Configuration for a Flussonic Server
Selecting a hardware configuration for a Flussonic server is a crucial step in ensuring that the system is able to handle the demands of streaming content and provide a reliable service for viewers. The process involves considering:
- Type of content that will be streamed.
- Expected number of viewers.
- Budget available for hardware.
In this article, we will outline the general guidelines involved in selecting a hardware configuration for a Flussonic server and provide some tips and considerations to keep in mind.
In order to select the most appropriate hardware configuration for your Flussonic system, you will need to consider the parameters of the entire network of Flussonic servers, including the number of subscribers, the number of TV channels in the lineup, and the parameters of the video stream (whether the stream is multibitrate or linear, the number of Mbps per video profile, the video and audio codec, etc.). You will also need to consider the type of archive services you will be offering, including the depth of the archive and the number of channels with DVR and catch up TV features. By knowing these parameters ahead of time, you can plan the minimum and maximum required capacity (in terms of channels, users, and network throughput) and select the hardware configuration for each type of server based on both economical and technical feasibility.
When selecting a hardware configuration for a Flussonic server, it is important to find a balance between the number of servers, their price and performance, and the cost of maintenance. While it may be tempting to use a single, high-performance server to handle the load, this can create a single point of failure and may not be the most cost-effective solution. The cost of hardware increases non-linearly with the performance, meaning that it can become exponentially more expensive as you try to increase the performance of the system. On the other hand, using too many servers can increase the cost of maintenance and take up more rack space. It is important to consider the reliability and redundancy requirements of the system and find a balance that meets both the performance and cost requirements. By spreading the load across multiple servers, it is possible to create a more reliable and cost-effective system that is able to handle the demands of streaming content.
For example, if you are an IPTV operator who already has a deployed base of set top boxes with certain capabilities, you will need to consider these capabilities (codecs, bitrate, etc.) when selecting hardware. On the other hand, if you are planning a greenfield deployment, you will need to consider the budget available for set top boxes and the available options and their capabilities. In either case, it is important to draw a preliminary network diagram with ingest, transcoding, archive, origin, and edge servers, and plan the min and max capacity of each type of server.
Testing the hardware performance
The first step in selecting a hardware configuration is to establish a baseline for the performance of the hardware using a scenario that closely imitates the production requirements. This can be done by setting up a server that will imitate the load and connecting it to the system being tested. Then, monitor the performance of the system closely, paying particular attention to the Erlang Scheduler metric, CPU and memory load, network throughput, disk system load, and Flussonic log file for any warning or error messages. You should also monitor the system temperature and power consumption. As you gradually increase the load on the server that is imitating the client connections, or as you add more channels to transcode for your Flussonic transcoder, you will be able to see at what point the system performance begins to degrade. This will allow you to identify any potential issues with the hardware configuration and adjust the parameters as needed.
It is important to note that Flussonic is CPU greedy, meaning that it will try to consume all available CPU cycles even if the system is not loaded. This is done in order to save time on allocating CPU resources when the load suddenly increases. As a result, the CPU load meter will always show that the CPU load is high, even if the system is not overloaded. Instead, pay more attention to the Erlang Scheduler metric, as this is a more accurate indicator of the system's performance.
We at Flussonic do not typically give specific recommendations on hardware configurations, as the same configuration that we test in our lab may perform differently on a client's site due to various factors such as misconfiguration, different versions of the operating system or Linux kernel, or the presence of other software services on the same server. However, we can provide some pointers to help clients get started, particularly for those looking for an optimal transcoder configuration. For example, we know that a NVIDIA RTX A4000 GPU is capable of transcoding approximately eight full HD streams into three profiles. Based on this information, we might suggest that a client start with a Dual Xeon Gold CPU, 64 Gb of RAM, and three NVIDIA RTX A4000 GPUs in order to transcode 24 streams in total. It is important for clients to test this configuration using their own streams in order to see if they can replicate the results.
It is worth noting that Flussonic also has its own lineup of hardware, including transcoders, which are actually appliance servers that we guarantee to deliver the performance specified. If you are interested in learning more about our hardware options, you can visit the Flussonic Coder web page at https://flussonic.com/hardware/coder/.
Overall, choosing a hardware configuration for a Flussonic server is an iterative process that requires careful planning and consideration. By following these guidelines and working closely with our technical support team, it is possible to select a hardware configuration that will ensure the success of your Flussonic system.
How to get help in hardware selection
In conclusion, the process of selecting a hardware configuration for a Flussonic server is a complex task that requires careful consideration of the type of content being streamed, the expected number of viewers, and the budget available for hardware. By establishing a baseline for the performance of the hardware, monitoring key performance metrics, and working closely with our technical support team, it is possible to select a hardware configuration that is both effective and cost-efficient, ensuring that the Flussonic system is able to handle the demands of streaming content and provide a reliable service for viewers. Whether you are an existing IPTV operator with a deployed base of set top boxes or a greenfield deployment, our technical support team is here to help you get started and ensure that your Flussonic system is configured and optimized for success.
At Flussonic, we understand that choosing the right hardware configuration can be a daunting task, which is why we are here to help. Our technical support team is here to assist you in pre-production phases, testing available equipment, making sure that it is configured properly, and optimizing its utilization. We will never leave our customers alone with the difficult task of planning the system, selecting hardware, and configuring Flussonic. However, it is ultimately the client who must make purchasing decisions, and Flussonic cannot and will not push them in any particular direction.
Ultimately, seeing is believing, and it is important to see for yourself how Flussonic works. Take whatever system is available in your lab, install Flussonic on it, and see how it performs. If needed, open support tickets and ask our technical specialists questions, or ask us to validate the Flussonic configuration and review logs if something goes wrong. This will help you establish the very first baseline and give you a good initial understanding of what the Flussonic system is capable of. With this knowledge, you will be well-equipped to make informed decisions about the hardware configuration of your Flussonic server.