White House says 'Obamacare' website will be fixed by end of November

Actually, during the early days when web server and web services are invented, sudden surge of web hits will cause congestion, overloading and web pages will fail to load and web service will malfunction - because the technology to handle web load and redistribute the load to other server resources is not available at that time.

However, given the current technologies in web load balancing - by using server load balancer (SLB) - to spread the load within server cluster within a data centre. As well as redistributing the load among varius data centres through data centre load balancing - to distribute the load between data centre via global load balancer - the multi-million of web hits should not be a problem to handle.

Moreover, with network able to handle gigabits network traffic and can be implemented to do "active-active" configuration via GLBP (Global Load Balancing Protocol) in the ingress port and BGP (Border Gateway Protocol) in the egress ports - load balancing of network traffic to redistribute the web traffic is possible.

Even though there are no tools to simulate the web load of multi-million web hits - for stress test - unless use DDOS (Distributed Denial of Service) method (hacking method) - by harnessing distributed computing to stress test by simulating multi-million hits, mental calculation will help.

E.g - simple math computation - like taking into account the capacity of a server resources such as processor, memory, harddisk, network capacity and speed - the number of web hit in which a server, a server cluster and how many data centres can handle, the multi-million web hits can be computed with very precise accuracy to determine the specs per server, per server cluster and how many data centres can be computed to handle the multi-million loads - without causing the problem in the program launch.

Thus the problem of web hits when multi-million concurrent hit the web - will not pose a problem - even through mental calculation without a load test.

Experience technical person handling such project will be quite confident such technical glitches - even of National scale will be mitigated and will not pose too much of a problem.

Lastly, the web service, application, program and database must be thoroughly tested to ensure they are bug free and can handle fast responses. The total response times that go through every devices and software should be aggregated to ensure that the web transaction can be completed within a designated time - even with concurrent access.

Many network and application management software that can measure the response time end to end from network, database, applications, programs, web services - up to code level can be deployed to do the actual measurement and if there are any issues, the network, system, application and web code and database - can be tweaked by changing the relevant parameter and fine-tune to ensure they can perform at a optimal level.

All likely issues can already have been ironed out - through the various tests before the final roll-out.

Thus an experience solution architect to oversee the project is required to ensure success for such national level project - and cheap cheap some more to employ.

 

(After note :- threshold on all the IT resources capacity can be monitored to ensure it can handle the traffic. If threshold is reached, more IT resources can be dynamically provisioned like CPU, memory, harddisk spaces, even network bandwidth through link aggregation.

 

eg. of SNMP configuration turn on for network devices to monitor the network devices capacity threshold:-

 

snmp-server community xxxxxx ro

snmp-server host 13.1.1.1 version 3 auth xxxxxx

snmp-server enable trap

 

eg. of dynamically provisioned additional bandwidth to network link using link aggregation :-

 

interface port-channel 1

switchport mode trunk

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 1,20,30,40,100,1002-1005

 

interface port-channel 2

switchport mode trunk

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 1,20,30,40,100,1002-1005

 

interface fa1/14

channel-group 1 mode active

switchport mode trunk

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 1,20,30,40,100,1002-1005

 

interface fa1/15

channel-group 2 mode active

switchport mode trunk

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 1,20,30,40,100,1002-1005