/s/blog _ 6582 aa 4 10 100 qd w8 . html
2. Formula for calculating the number of concurrent users
Number of system users: the number of users rated by the system, such as an OA system, the total number of users who may use the system is 2000, so this number is the number of system users.
Number of people online at the same time: the maximum number of people online at the same time within a certain time range.
Calculation of average concurrent users:
C=nL / T
Where c is the average number of concurrent users, n is the average number of users visited every day, l is the average time from login to logout in one day (average operation time), and t is the length of investigation time (the length of time users use the system in one day).
Calculation of peak number of concurrent users:
C is approximately equal to C+3* radical C.
Where c is the peak value of concurrent users and c is the average value of concurrent users, and the formula follows Poisson distribution theory.
/kloc-What is the possible number of concurrency caused by 0/00000 users? Solution (theory)
/info-detail-6 1758.html
Since there are these two formulas, we assume the possible concurrency of 6.5438+million users.
1.n Number of users visited every day =100000.
2. Suppose the service is used for online banking, L= the average time for users to log in and log out in one day is set to (5 minutes), t suppose users visit at 8:00- 12 every morning. Duration 16 hour, or 960 minutes.
(This number of users is assumed to be the average number of users accessing the system every day. If it is the total number of users, then we need to calculate the average number of users who visit the system every day. )
C =10000000 * 5/960 = 52083.33/m (i.e. 52083.33 per minute)
3. The peak of concurrent users is
C'≈52083.33+3 * radical number 52083.33 = 52083.33+3 * 228.22 = 52767
It feels a little strange. Maybe some of my parameters are set unreasonably. Perhaps the calculation of these concurrent numbers should be divided into busy hours and idle hours instead of days, which may be more accurate. Anyway, let's discuss the hypothesis according to this idea first.
However, it should be emphasized that in the information I found online, some calculations are based on hours and some are based on minutes. I use minutes to calculate here, so I think the average number of concurrent users should be in time.