Last week, the SAN server went down and the team struggled for hours to get anyt
ID: 3891815 • Letter: L
Question
Last week, the SAN server went down and the team struggled for hours to get anything done. They were literally passing USB drives to each other to share data. This is not the only server that experienced an outage! The day before, a customer-facing Windows server used to place orders was down as well for an hour during peak ordering time and people were writing out the orders by hand on paper and faxing them back to the customers. You know that there must be a better way!
1.What solution do you suggest?
2.How do you keep the costs down?
3.How can you implement it without very much disruption to the business?
Make sure your solution is global and is not limited to just these two servers.
Explanation / Answer
To give a proper solution one must first understand the problems. In case of an SAN server, is that it is a network that provides a consolidated block level data satorage.They store the data over storage devices like tape libraries,etc.However the only problem with this server is that if this goes down the entire network goes down.As there is no backup for the network stored data anywhere all the data operations are forced to stop.
In case of an windows server the main problem is that if there is a single server running for multiple machines then the server tends to slack down due to huge incoming traffics. As such when the traffic buffer is more than its throughput it stresses and thus crashes.Thus when the only customer facing server crashes down, the entire operations stops.
1. The main solution to this problems is using a using a Distributed - Cloud based Network server.In a distributed cloud based server not all the data is stored at a particular network drive.Instead chunks of data are replicated and stored in multiple locations. As such, even if a particular device becomes offline ( due to crashes or network overload ) there tends to be no problems as the data is retreived from another alternative node (device).Moreover this cloud based distributed devices are known to keep a backup copy , which can be used for disaster managements.This process is being extensively used by carge corporations and banks. As such this process will surely be helpful.
2. To keep the costs down in cloud based distributed systems the following principles may be used-
a)Instead of diverting all the traffic to a single node(device) divide the traffic to multiple nodes.As such the load over a particular node will be less and there will be lesser chances of a breakdown
b)Always use a cloud management service that tracks down the resources being used and the resources not being used.This helps to keep a good tack of the pricings of a cloud platform and cut down the costs by eleminating the unwanted resources.
c)Use auto-scaling in servers.Auto scaling automatically scales down the work of an server then there is low traffic as suc lower resources is used and lower costs.
d)Eleminate empty storages.Find all the resources that are not being used but in storage and force them to be flushed out after a given stipulated time.A simple script in cloud should to the trick.This will reduce the costs and hence reduce the costs.
3) The server can be easily moved from SAN to cloud without disrupting the business.First we need to identify the cloud service that we are going to use.There are many distributed cloud solutions namely Amazon AWS,Microsoft Azure,etc.These are some of the prominent services.
The next step is identifying how much storage we need in the cloud and thus configuring the cloud server.This is quite easy as most often these platforms comes with an GUI that lets you to do the hard stuff.And moreover the stoprage needed in cloud is equal to the storage needed in SAN.(To be safe include a cloud management program along with this).
Now the next step may be done in the following steps :
* Privately connect your SAN with the cloud service.
* Privately sent petabytes of data from SAN to cloud over a secured network
* Build a hybrid storage that preseves the state of the migrated data and create a cloud service accordingly.
* Send the data in batches from hybrid storage to cloud.
* Once all the data is sent over to cloud test your data services and your business is ready to work with.
Since all the data transfer is being done over a secured private network and in batches,this is in no way going to sterss the SAN we are working with as such, we can work with the san while the business is online.
This is a global solution to all types of networks,
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.