One of this first questions we asked when deciding to sue Amazon’s AWS services was: Will we save money?
At first, your EC2 servers may be simply architected, perhaps small instances. Once a more serious commitment is made for a more robust architecture at Amazon, there will be additional costs introduced. For example, a small instance, running a 2 disk 100GB EBS volume w/RAID0 plus monitoring and a static IP, and the cost goes from the current ~$68 monthly per system to ~$88 monthly per system. Bump the instances to large instances (likely needed for any server running mysql, or any other equally intensive application) and that cost goes to ~$265 per instance. Add in the costs of the additional services like bandwidth, Static IP, Cloudwatch etc and the costs can quickly escalate. Of course, upfront payments for Reserved Instances can drastically reduce the costs further.
However, the savings in development and deployment costs I think far outweighs a narrower gap in the savings between physical and AWS servers, and the real MRC on the AWS servers will likely be lower for a given amount of computing resources.
So- can you save money? Yes. In some cases, it will be a direct apples to oranges savings of hard dollars. In other cases, the agility gained will provide the greatest savings. In most cases- a combination of both will drive your cost savings.
To learn more about how operating in the cloud can save your company money, contact us for a free consultation.
I recently saw the headline,”T-Mobile and Microsoft/Danger data loss is bad for the cloud“, and, as an admin who works with cloud technology on a daily basis, viewed the headline with some concern. However, after reading the article itself, my only thought is “What does this have to do with the cloud?”. Reading through, we find that Microsoft/Danger stores your phone data (contacts, photos, etc) on it’s servers, and that the phone needs to constantly be in contact with the servers in order to maintain service and data. Unfortunately, the servers crashed, and all of the data was lost. Turn off your phone, lose all your data. Yet- this is exactly what the Sidekick service promises to protect you from- and it failed.
The problem with blaming this on the “cloud” is that, while technically, your cell phone and the Microsoft/Danger servers form a “cloud”, the failure lies with the servers, and those who administer those servers. It doesn’t matter whether those servers are virtual, or physical- if there is not a disaster recovery plan in place, and if that disaster recovery plan has not been tested- data will be lost. Your data. This is not a shortcoming of cloud computing- it is a result of depending on others to maintain your data. It also gives us caution when depending on external providers over the network to always be available. Services stop. Power fails. Disks die. Routing interruptions happen.
This is just network computing. But if the people (or companies) behind it all don’t do their own due diligence- disasters like the this, and worse will continue to happen.
Bio-Linux and other bioinformatics tools available for EC2, Amazon’s Elastic Compute Cloud, were recently highlighted on the Amazon Web Services (AWS) blog. Customized Amazon machine images (AMIs) allow for the packaging and rapid, web based deployment of the data sets and tools needed for these specialized tasks. Because AMIs can be saved, reproducing past results is simplified and because these can also be shared, the computation environment of a particular analysis can be easily replicated both from within and outside your organization.
Continue reading “AMIs for Bioinformatics on AWS”
Yesterday’s trip to Cloud Camp Boston was most interesting. The keynote gave a good overview of what cloud computing is, where it came from and where it is headed. Continue reading “From the Cloud Camp”