In my recent blog on how virtualization is the building block of cloud computing, I was asked a question about how load balancing and redundancy work if a server in the cloud crashes. This is an excellent question -- and a useful one in seeing how cloud suppliers can provide high levels of uptime while running commodity hardware.
A load balancer is a physical device or machine running in the cloud that receives requests from clients to supply a Website or application running in the cloud and must choose a server in the cloud running the requested Website or application. Most clients have multiple virtual machines running on several servers for the load balancer to choose from. Users want to avoid having all the virtual machines running on one server, because they would lose all access should that server crash. The load balancer would look for the server with the least amount of CPU or memory use (settings determined by administrator) and select a VM on that machine to ensure the best performance for the client from a given server.
The image below from Cavemans's blog shows the load balancing concept with the addition of a firewall in front of the load balancer.
A Firewall in Front of a Load Balancer
As for redundancy, you can easily see how a set of virtual machines being load balanced across multiple servers is always available, regardless of the status of a given server. Should one of the servers go down, clients will still have access to the VM. It will be available on other servers and can be started on new servers easily to keep the number of available VMs constant.
A good, simple overview of load balancing in the cloud is presented in this four-minute video. The presenters use the term "horizontal scaling" to describe the results of a load balanced set of virtual machines.
Stephen – those are good drivers for Cloud but I also believe that access to data from various platforms and from anywhere is another. Using the cloud as a centralized repository is one area that seems to be gaining the most acceptance. Teams collaborating to develop an end product that is built up from many elements. A project that I am currently working on uses a Cloud solution. The architect drops in the CAD files and then we can add in the appropriate layers to the file and upload the modified CAD file.Once approved the installation teams down load the approved documents and use them to build out the space.Each is done with permissions and approvals but this is a great way to collaborate.
I've always thought that the superior load balancing capabilities of the cloud (particularly the public cloud) is one of its major selling points. Question for Stephen and the group: is load balancing tempting enough on its own to get an organization to move to the cloud, or are its other potential benefits more important to the to-cloud-or-not-to-cloud decision?
Ashish - Couldn't have said it better myself. I am currently listening to the Dell application virtualization webcaste. and at the same time preparing do to a network assesment that will prepare a site for voice virtualization and data virtualization.
Technocrat - I resently hosted a Health Care breifing at Polycom. Our own Dave Wangner was able to partcipate and he had some great insight. Nice to be able to share with the group and continue to expand our horizons.
The blogs and comments posted on EnterpriseEfficiency.com do not reflect the views of TechWeb, EnterpriseEfficiency.com, or its sponsors. EnterpriseEfficiency.com, TechWeb, and its sponsors do not assume responsibility for any comments, claims, or opinions made by authors and bloggers. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
4/29/2014 - Join Dell and Intel for an interactive discussion about implementing, refining and improving your virtual environment. Specifically we’ll discuss pain points virtualization can solve and those that it can create and how to prevent them.
Enterprise Efficiency is looking for engaged readers to moderate the message boards on this site. Engage in high-IQ conversations with IT industry leaders; earn kudos and perks. Interested? E-mail: email@example.com
Dell's Efficiency Modeling Tool The major problem facing the CIO is how to measure the effectiveness of the IT department. Learn how Dell’s Efficiency Modeling Tool gives the CIO two clear, powerful numbers: Efficiency Quotient and Impact Quotient. These numbers can be transforma¬tive not only to the department, but to the entire enterprise. Read the full report
Now that TGen has broken new ground in genomic research by using Dell's storage, cloud, and high-performance computing solutions, the company discusses what will come next for it and for personalized medicine.
The Translational Genomics Research Institute wanted to save lives, but its efforts were hobbled by immense computing challenges related to collecting, processing, sharing, and storing enormous amounts of data.
Office and personal productivity tools come in a first-class and coach flavor set, but what makes the difference is primarily little things that most users won't encounter. What's the big issue in using something other than Office, and can you get around it?
We really don't want an "Internet of Everything" but even building an Internet of Everythinguseful means setting some ground rules to insure there's value in the process and that costs and risks are minimized.
Google's Chrome OS has a lot of potential value and a lot of recent press, but it still needs something to make it more than a thin client. It needs cloud integration, it needs extended APIs via web services, and it needs to suck it up and support a hard drive.
On a recent African trip I saw examples of the value of the cloud in developing nations, for educational and community development programs. We could build on this, but not only in developing economies, because these same programs are often under-supported even in first-world countries.
VMware's debate with Cisco on SDN might finally create a fusion between an SDN view that's all about software and another that's all about network equipment. That would be good for every enterprise considering the cloud and SDN.
Wearing a bulky, oversized watch is good training for the next phase in wristwatches: the Internet-enabled, connected watch. Why the smartphone-tethered connected watch makes sense, plus Ivan demos an entirely new concept for the "smart watch."
Cloud storage costs are determined primarily by the rate at which files are changed and the possibility of concurrent access/update. If you can structure your storage use to optimize these factors you can cut costs, perhaps to zero.
The Internet has evolved into a machine for drumming up a chorus of "Happy Birthday" messages, from family, friends, friends of friends who you added on Facebook, random people that you circled on G+, and increasingly, automated bots. Enough already.