Enter your email address:
Delivered by FeedBurner
Previous Blog Posts
Wednesday, May 30, 2012
Guest post by Jeanna James
More and more I find myself on customer calls where a client wants to move 50TB or so of data to the cloud. In theory, moving data to the cloud should be straightforward, but in reality there are important issues to consider first. Reality, of course, consists of lots of ones, zeros, and some basic math. I'm no math whiz, so I decided to write this blog to help others like myself calculate the cloud equation. When talking with these clients, I ask the following questions: How much bandwidth do you have? How much data do you have? Do you ever want to restore your data? How long is acceptable for the restore process to take place? Does your cloud provider allow you to seed the cloud (i.e. an option whereby a disk drive or other appliance is sent to customers, who back up their data locally and then ship the drive/appliance back to the provider)? Does your cloud provider have servers in the cloud so you could potentially restore data directly from the cloud storage?
Let's walk through two examples of customers planning to move 50TB of data to the cloud. One customer has a T1 line and the other is blessed with a fiber optic OC3 network. When I tell someone writing data into the cloud can take months, I'm not joking. I'm also not good at doing spontaneous math over the phone or on the white board. I should also confess right now that this is a three part blog and in this part we're only going to cover one solution to the problem I'm about to expose.
Example 1: Customer with T1
Example 2: Customer with OC3
Now for more fun facts about the above calculations – they assume a pristine environment with minimal WAN overhead, but no other network load. What, you use those lines for Internet, email, and IP phones? Well, you can see where that is going to be a problem.
Below is an eye chart that breaks down 10TB of data based on these same assumptions.
|Type||Effective Bandwidth in Megabits Sec||Effective MB Sec including Protocol (MB/s)||Time to Transfer 10 TBTBTBTBs|
So why am I, a senior manager of cloud business development, being a Debbie Downer and pointing out these cloud-based challenges? Writing data into the cloud is just one step. Getting the data back, especially if it is backup or archive data is even more critical if the cloud is going to be the storage target for your company in the event of a disaster. This drives home the importance of other questions I ask about the cloud providers our customers are considering: Does your cloud provider allow you to seed the cloud? Does your cloud provider have servers in the cloud so you could potentially restore data directly from the cloud storage? If you have a disaster, will your cloud provider ship hard drives with your data back to your site for recovery? The answers to these questions are critical.
With the advent of massive, 50TB cloud solutions, how are customers overcoming the above math challenge? Today we'll cover one potential solution to this problem. In this example, the customer is working with a cloud provider who: 1) allows customers to seed data in the cloud and 2) bases their compute cloud on VMware in order to take full advantage of physical-to-virtual (P2V) technologies. There are several advantages to this type of cloud solution.
First, seeding the cloud gives customers the ability to ship media to the cloud provider and then merely send changes over the wire. This significantly reduces the bandwidth required by leveraging integrated data reduction technologies such as source-side deduplication and compression.
Second, in the event of a disaster, the customer has the ability to perform recoveries into the compute side of the service providers cloud. Not only can they perform these recoveries, but because the service provider is running VMware, they can leverage P2V technologies – in CommVault's case our Virtualize Me functionality, to automatically recover a physical server into a VM. In addition, organizations should ask their cloud provider if they can access tools to multi-stream data to the cloud and maximize bandwidth, especially for those customers who have a dedicated 1Mbps pipe. CommVault provides this feature in Simpana 9. This is one scenario that will substantially accelerate an organization seeking to move large volumes of back up or archive data into the cloud. I've heard from customers who use modern data management features such as dedupe, compression and multi-streaming to cut the time it takes them to move large volumes to the cloud from months to days or even hours.
Planning such an implementation is critical to the success of the project both for getting data into the cloud and for testing restore objectives. In my next post we'll look at shipping hardware and when physically shipping a device might be the best option for a customer.
Learn more by watching our Virtualize Me demo on YouTube.
Jeanna James is senior manager of cloud business development at CommVault.
Please note that your comments will be sent directly to the author of this blog, and will be published upon approval per CommVault's comment policy.
The content of this blog reflects the thoughts and opinions of the author, and does not represent the thoughts, opinions, plans or strategies of CommVault Systems, Inc. ("CommVault") and CommVault undertakes no obligation to update, correct or modify any statements made by the author of this blog. Any and all third party links provided by this blog are not affiliated with, nor endorsed by, CommVault.