Did we need separate link for very important systems

We have a network compined of LANs and WANs acros many cities. We are using Frame relay in WANs. We have alot of systems that uses this network such as:

- Many client server applications using many MS-SQL servers
- Exchange server and Outlook clients
- Few FTP servers
- Share folders

Our main problems in this network:

- High traffic
- Low band widths in some WANs
- Network disconnection some times

Our systems needs continues connection all the time, although the data transfered between client and servers is not too much, but it is very frequent transfer, so that network disconnection is the most important problem.
What did you suggest as a solution for such problems.

Did we should think to separate special link for important systems, or trying to solve the problem for all the network?

Thanks for your patience.
Motaz Abdel Azeem
Who is Participating?
ITsheresomewhereConnect With a Mentor Commented:
In your original Question you stated three problems 1) high traffic 2)low bandwidth 3)constant connectivity.  You also restated the importance of "needing constant connectivity".  Therefore you have completed, although roughly, part of #1, define priorities  But the order of importance is:

1. Constant connectivity throughout the network
2. Manage bandwidth availability at all points to attain goal number 1 through revision of service or introduction of alternate services
3. Monitor traffic and services and take appropriate actions to adjust network availability to maintain the primary goal.

What your discussing then is what would be called a network design analysis, or a network engineering review, or a network infrastructure review.

It is a big task, and it can be very complex when multiple offices, servers, and telecommunication services are involved.  To be done well, and to have lasting meaning, it generally requires a strong understanding of business model and needs, network engineering and design standards, knowledge of available solutions, proper monitoring and analysis of traffic and tools to do such with.  This is not a one man project, it does take time and it will cost money.  Depending on how urgent or how critical the problem is, or if there are limited skills within the company, then outside services should be called upon to perform either all or various portions of the project.  

Companies that specialize in network design engineering, systems integration, or telecommunications design engineering can provide these types of services, possibly at minimal charge to you, as part of other services you are already purchasing from them.  By that I mean, whom ever provides your WAN may have engineering assistance available to review and recommend changes.  If not the company that competes against them probably does and might give you the services as a carrot to switch your business to them.  I have had lots of "free engineering" done as a result of telling competitors that I am unhappy and may change my business.

Now the study is going to take into account 2 primary factors:
1) what is the physical layout of everything on the network, the MAP, as sysExpert stated, and if you don't have at least a partial one then you need to build one.  

How?, a really big map or big sheet of paper to start.  Mark every location and all the types of equipment, servers, terminals,etc that are in use at each one.  Also you connect those dots with lines as to the type of data communication link you have between them all.  Make it as detailed as possible, because one thing you might find is you have the same size link for a office with 200 people that you have for 20.  Not good  either too much or too little depending on its size.  

So you need the MAP the picture the ability to look on the wall and say this is connect by this to this and this is where it is breaking down.

2) The other side of the design is how does the information flow, is it local, when does it go to central servers, or does it have to go all around the network just to get an answer.  Knowing how the information flows then can lead you to what changes might need to be made to balance your network for good response and information being available.  It will also point out where the "critical" links are and which ones should have more then one way to get there, the redundancy.

Now when you have the map and know how the information flows you can then start putting up a problem list for each specific site or link or reported problem.  That way you address the ones with the most problems or are the most crucial and in turn take stress off the network, thus improving the whole.  You also see when you will have to improve things because you will start seeing patterns earlier.

Simple analogy you can't eat the whole dinner in one gulp, you take the parts and enjoy the meal.

Not all problems will be resolved by changes to communication links, some may involve where information is stored, and some may be changes to hardware, bigger pipes connected to small faucets don't solve the problem.  But the plan leads you to the solutions.

Hope this has been of help.


1) Map out your network and define tour priorities.

2) Define the setup so that you have backup lines for critical operations. These should failover automatically.

3) Monitor all the lines for usage and decide what sections need to be improved, perhaps using compression or a VPN.

I hope this helps !
MotazAuthor Commented:
Thanks SysExpert, but how can I do step 1 and 2?
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

>What did you suggest as a solution for such problems

Hire a good network consultant to come in and baseline your network and make recommendations.
MotazAuthor Commented:
As programmers, what kind of help can we give the network designers?
btw. Network engineers and frame relay engineers are part of our IT department.
Well I imagine your assistance or contributions could
fall into the areas of

1. knowledge of the business model and how its needs have been addressed through programming.

2. impact of applications on load factors and how changes might be made to reduce same though programming solutions.  

3. construction of monitoring assists or addins which provide information regarding trouble spots, warnings etc.

4. knowledge of the data structure, efficiency of databases, data access issues and solutions, and flow through out the network.

Generally the contribution of knowledge to the review and the construction of tools towards solutions, along with willingness to introduce newer lighter weight more efficient programming solutions which make for a more robust network.

Good code, stable code, secure code that executes smoothly and quickly.   Windows is probably not a good role model.
For 1) You can use a network mapping tool like Visio Enterprise, or any of the other mapping tools.
Also to see usage, install the free MRTG and monitor the routers.

2) Management has to sit down with you to decide what is mission critical,and what can be used alternatively.
You need to figure out the cost benefits of having backups in case of a failure, and see what it costs to be WITHOUT a backup connection. This will give you the financial backing for your project.

I hope this helps !
MotazAuthor Commented:
Thanks for your time and help
Glad to have been of assistance.  Good luck on your project.

As you go along remember Experts-Exchange is always open.

Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.