business continuity using NAT for connection to remote server

0 pts.
Tags:
Availability
Bandwidth
Networking
We have two servers working on hot availability mode. They are doing ticketing transaction. Is there a way to provide business contiuity by automatically connecting the users to a remote server. The packets from users should get diverted to the new server using NAT to be implemented with a cisco router. Replication will be used to keep the remote server uptodate. How do we implement a time delay for starting the NAT(delay from the time the router senses that the servers are down) and what about restoring back. Generally speaking, is this idea all right or fraught with problems which I am not able to see.
ASKED: October 3, 2005  10:15 PM
UPDATED: October 5, 2005  10:59 PM

Answer Wiki

Thanks. We'll let you know when a new response is added.

I think the main issue here is you are doing transactions. Any failover would have to accomodate the state of the transaction. From what I have read, the most common model for what you are doing is to have parallel front end web servers providing a customer interface, and a back end SQL server cluster doing the actual transaction.
This is much more invloved than just redirecting traffic when a server fails.
rt

Discuss This Question: 3  Replies

 
There was an error processing your information. Please try again later.
Thanks. We'll let you know when a new response is added.
Send me notifications when members answer or reply to this question.

REGISTER or login:

Forgot Password?
By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy
  • Tvenkat
    you are correct. The failover solution today is provided by two servers working on High Availability mode. In case of a failure of a server, the IP of the failed server is acquired by the second server and as both servers access the same common storage, transaction keeps on going. What do we do if both the servers conk off. Such instances do happen from reasons not directly connected with server hardware. (say from UPS problem and so on). In such a case how do we use a remote server to keep the transactions going? Definetly users will have to login again and transactions which are cut off midway have to be recovered. But still we can continue our business. In this context, can we use NAT for automatically and more importantly transperently connect users to new server?
    0 pointsBadges:
    report
  • Astronomer
    Why don't you just extend your cluster to more nodes? Have you investigated majority node clusters? This would allow SQL to fail over more than two nodes. Microsoft even says the nodes can be in different campuses from each other. There shouldn't be an issue adding more nodes to the web front end. As far as failures like UPSs are concerned, do you have servers with multiple power supplies? If so, you should plug them into different UPSs which in turn are plugged into different circuit breakers. When I was asked about how to power our new rack, I recommended we use UPSs in pairs and wire them into separate legs of the 220 coming in. I would be surprised if current technology didn't allow you to have two campuses linked so if one campus goes down, the other would take over so your business could continue. I don't expect this to be cheap, but it is doable. You would need to consider all critical services. This gets complicated fast. What about your ISP? A failure here could stop your business. If you ISP says this won't happen, do a web search on the phrase "backhoe fade". When I set up a small ISP a few years ago, The local phone company, (who saw us as competition), "accidentally" cut our T1 four times. Our local PBS station expressed the need to be on the internet no matter what, so I am planning a BGP solution to allow failover to another ISP. What about power? Do you have two locations separated enough to minimize the possibility of power going out at both places at once? The TV station has standby generators so they can be stand alone in under 15 minutes. This should give you enough of an idea of what is possible. You can buy failover protection. It's just a case of how bad do you need it, what level do you want, what can you afford, how much effort are you willing to invest. My suggestion for what you have described is to implement clusters across more nodes. This way you can keep state with customer transactions and continue your business. rt
    15 pointsBadges:
    report
  • Tvenkat
    thanks for your time and effort in giving a detailed suggestion. I am just looking for a solution implementable within the resources available. I agree with you that the 'whole' picture requires a comprehensive approach. Although microsoft may claim that the cluster works across campuses, my understanding is that there is a limitation of around 20 km (this was the case with sun a few years back). The problem is in the latency created by distance. we are looking at a remote server(already available with spare capacity 500 km away) Checked out on backhoe fade. very interesting. came across a humorous post at link given below http://www.petting-zoo.net/~deadbeef/archive/745.html
    0 pointsBadges:
    report

Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to:

To follow this tag...

There was an error processing your information. Please try again later.

REGISTER or login:

Forgot Password?
By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Thanks! We'll email you when relevant content is added and updated.

Following