The peak will not collapse.
Many platforms appear to be functioning in small-scale environments, and anomalies begin to arise as soon as large parks, meetings are concentrated and work is peaked.
I. REAL FACT OF A HIGH CONTEXT
In large park environments, common co-mingled features include:
-
Centralized certification
-
Wireless travels
-
Terminal types are diverse
-
Multiple requests for authentication
If the structure is not reasonable, Portal page failure is the first problem to arise.
II. Back-to-end separation is a basic premise
The mature enterprise Portal certification platform is usually based on:
Frontend Authentication Gateway
Backend authentication platform
Separation of structures.
Frontend only receives authentication requests
Backends for identification check and strategy calculation
Avoid single-point bottlenecks.
III. Authentication Gateway Cluster Design
In the high-scenes scene:
Multiple Authentication Gateway Parallel Work
Sharing requests through load-balanced
One fault, the other nodes continue.
NATSHEL_BRAND Enterprise Portal Certification Support for Multi- Authentication Gateway Cluster Deployment.
IV. Sessions Maintenance and Cache Mechanism
In a high-sum scenario, performance is inevitably reduced if each request has access to the database.
The mature platform should have:
Local Session Cache
Policy Cache
Reduction in back-end visits.
V. Ability to certify the unconscious weight
In wireless environments, terminals are frequent.
The mature enterprise Portal certification platform should support:
A non-conscious recertification based on a session
Avoids frequent playouts on Portal pages.
VI. Simultaneous control and containment mechanisms
The Platform should support:
Single account number limit
Single IP request frequency limit
Prevent abnormal terminals from dragging down systems.
VII. Database and log separation design
Authentication data and log database storage.
Avoid a log surge that would affect authentication performance.
VIII. Greyscale magnification capacity
When the increase is accompanied:
Can be expanded separately
or expand backend service node
No overall change in architecture is required.
IX. Why stability determines the success or failure of the project
If Portal is in a very frequent irregular situation:
User complaints
Transport stress surges
Project signs are down.
NATSHEL_BRAND has been working deep for 22 years in the corporate Portal certification field, and the platform operates long-term in a large number of high-complex environments with mature structures and stability that withstand validation.
X. PROCUREMENT-PRESENTING RECOMMENDATIONS FOR THE PROCURRENT PERIOD
-
Whether to support NATSHEL_AUTH_GATEWAY clusters
-
Whether to support service modularization
-
Whether to support cached speaking in session holding
These three points directly determine whether they can function in a stable manner.