Scaling your dynamic services in a corporate environment or cloud

8:27 PM
Scaling your dynamic services in a corporate environment or cloud -

When we refer to large data centers and Cloud, scalability comes to mind and becomes a concern. ADCs and load balancers are critical in such deployments. Although the burden of these services goes UP due to the demand of end users, service owners bring new entities and resources to meet the need. These new entities need to have their definition in the load balancer for them to participate in the transformation of the actual traffic.

Couple releases NetScaler had introduced the service concept based on the area where the service definition on NetScaler is done through the DNS response for the domain. Similarly Domain Services group based has been introduced which consists of members whose IP addresses are obtained by solving the domain names of the servers related to service group. name resolution process may return more IP addresses for the same resource, but we consider that the first IP returned and ignore the rest.

NetScaler 10 , which is a milestone version for all types of cloud deployment, introduced AutoScale option for service groups on the domain basis. You can set up a group-wide based domain services automatically on the basis of all the IP addresses returned by DNS response. It is added as fast enable / disable the option that gives you the option of automatic scaling.

Once a domain based server is linked to the service group with AutoScale enabled, NetScaler dynamically creates and binds to TCP and UDP monitor in the service. NetScaler default uses UDP monitors to send DNS queries to the name server. If the DNS response is truncated NetScaler then falls to dynamically monitor and TCP uses TCP as a means underlying for exchange of DNS messages ensuring that we receive all the data sent by the server. Once NetScaler starts using TCP Monitor does not go back to the UDP screen in real time.

When NetScaler receives multiple IP addresses as part of the single DNS response, it queries all the IP addresses individually and followed up by default. NetScaler ensures that only adds multiple IP addresses to the service group if UP and answer monitor query. Overall NetScaler takes care of setting up the infrastructure with caution, ensuring the provision of good service. Now the question is how we manage these deployments in real time because the DNS records expire after associated TTL ??

Hmmm ... interesting and we'll see what's best NetScaler can do here ... once the registration expires, the NetScaler associated UDP or TCP Monitor uses for sending the resolution request on the same server. After receiving the answer that valid for any changes. In case of no change, the roll continues to operate as such. If new IP addresses found in the response NetScaler acts Scale Up role and made the same set of validations for new IP addresses. After the validation passes, IP addresses are added to the service group for the treatment of traffic. If the DNS response was not specific IP addresses that were received last NetScaler then played very well scale down role and deletes IP addresses from the definition of a service group . IP addresses of withdrawal could be difficult as it may be in direct transactions and connections. NetScaler is the graceful closing connections to these IP addresses and does not end abruptly by the bad experience of the end user. In this period NetScaler ensures that no new connection and the transaction is authorized to be served by these IP addresses and only those existing gracefully terminate.

This kind of integrity in the handling of very special scalability NetScaler for such deployments .... What else ?? ... It is a little more NetScaler is to provide sufficient flexibility with your deployment. You can include static server based on IP and domain based together in the single service group. This gives you flexibility at some point in time you have static members that serve customers. This is important from the standpoint of availability as single NXDOMAIN response may result in removing all associate members in this area, resulting in complete failure. So a mixed mode approach might be best to have. Interesting ... and what more ?? You are also allowed to have multiple domain servers with AutoScale ON in service based on a single domain .... Now you need to think about all the use cases that use scarce as before the ... :)

Previous
Next Post »
0 Komentar