Capacity specifications for multimedia messaging

Last Updated : Feb 22, 2021 |
Prolog information
Metadata
Keywords:
The following sections describe the maximum capacities for Presence Services deployment topologies to help you calculate the best message persistence time (PT) for the various topologies.
Maximum capacities for Presence Services deployment topologies
Topology
Maximum number of users
Umax
Maximum number of persisted messages
Mmax (millions)
Description
Cluster of 1 VM-Profile-3 node
Up to 1,000
Up to 0.9
Non-HA topology
Cluster of 1 VM-Profile-4 node
Up to 5,000
Up to 4.3
Cluster of 1 VM-Profile-5 node
Up to 11,000
Up to 9.5
Cluster of 2 VM-Profile-5 nodes
Up to 11,000
Up to 9.5
HA topology
Cluster of 3 VM-Profile-5 nodes
Up to 22,000
Up to 18.8
Cluster of 4 VM-Profile-5 nodes
Up to 33,000
Up to 28.3
Cluster of 5 VM-Profile-5 nodes
Up to 44,000
Up to 37.7
Cluster of 6 VM-Profile-5 nodes
Up to 55,000
Up to 47.1
Cluster of 7 VM-Profile-5 nodes
Up to 65,000
Up to 55.7
Cluster of 8 VM-Profile-5 nodes
Up to 75,000
Up to 64.3
Cluster of 9 VM-Profile-5 nodes
Up to 84,000
Up to 72.0
Note:
When the maximum threshold for message capacity is reached (which is 95%), the system deletes the oldest messages until the lower threshold for message capacity is reached (which is 92%).

Persistence Time (PT) of a message

Use the following algorithm to calculate the persistence time of a message using the preceding table for maximum capacities for Presence Services deployment topologies. The advantage of this algorithm is that it is not based on traffic patterns.
  1. Determine the number of users (Ua) and the average message rate per user (MRPU) for your topology.
  2. Select a topology from the preceding table that closely matches your topology such that Umax ≥ Ua. Note the corresponding value of Mmax.
  3. Calculate the message persistence time using the formula: PT = Mmax / (Ua x MRPU).
  4. Verify if the persistence time matches your requirement. If you need a longer message persistence time, select a topology that is next in the row in the table.
Example:
The following example assumes a HA topology set up with 10,000 users, requiring an average message rate of 200 messages a week for each user. Using the algorithm, you can calculate the persistence time of a message as follows:
  1. Determine Ua and MRPU: According to the requirement, Ua is 10,000 users and MRPU is 200.
  2. Select a topology from the maximum capacities table: The most economical topology is a "Cluster of two VM-Profile-5 nodes". This topology has Ua of 11,000 users, which is greater than required 10,000 users. Also, the corresponding Mmax value is 9.5 million.
  3. Calculate the message persistence time: PT = Mmax / (Ua x MRPU) = 9.5 million / (10,000 x 200) = 4.75 weeks.
  4. Verify if the persistence time matches your requirement: If, for example, you want a longer persistence time, choose a topology that is in the next row of the capacity table, which is, a "Cluster of 3 VM-Profile-5 nodes". The corresponding Mmax value is 18.8 million.
    The persistence increases as follows: PT = Mmax / (Ua x MRPU) = 18.8 million / (10,000 x 200) = 9.4 weeks.
  5. Repeat the calculation until you are satisfied with the message persistence time.