It is no secret that organizations are increasingly looking into hybrid cloud solutions to get the best of both private and public clouds. And judging by recent announcements, service providers are paying attention. Last week Amazon augmented their existing AWS offering with a new desktop-as-a-service, while VMware this year launched its vCloud Hybrid Service (vCHS) as well announced the acquisition of Desktone, a leader in desktop-as-a-service.
Making the decision to migrate your VMs from private cloud to a public cloud involves many factors such as cost, performance, utilization, and sizing. But even just understanding whether or not it’s economical to run a particular set of VMs in a public cloud can be difficult. To get a cost estimate you have to manually import details of all the VMs and its resource consumption and then map the details onto the pricing model of the cloud service provider(s) you’re looking at. Depending on how how many VMs you’re looking at and how much you like Microsoft Excel you are looking at days to weeks worth of work.
Introducing CloudPhysics Cost Calculators for Hybrid Clouds
CloudPhysics now offers a way for administrators and managers to get pricing estimate for hybrid clouds in just seconds. We do the data collection, private-to-public cloud VM mapping, and cost calculation for you. You can even experiment with parameters that impact pricing for the target public cloud to figure out the best fit. This information is available for any part of the inventory: an entire vCenter, a single cluster, a single datastore, even down to a single virtual machine.
(If you’re not already using CloudPhysics, it takes less than 5 minutes to sign up and get started with our free trial and use these cards!)
In this post, I’m going to dig into into the Cost Calculator for AWS and I’ll talk about vCHS and doing comparisons across clouds in my next two posts.
Cost Calculator for AWS
Matching your existing VMs to one of the available EC2 instance types is the first step toward understanding what it might cost to run that VM in EC2.
Amazon provides EC2 reserved instances and spot instances. EC2 reserved instances provide fixed and discounted hourly pricing based on reserving those instances for a fee ahead of time. However, pricing of Amazon EC2 reserved instances depends on a number of factors such as instance type (which come with pre-defined sizes and performance capabilities), resource requirements, expected utilization, location, reservation length, type of storage, type of the guest operating system and more.
You need to understand:
Which EC2 instance types are the right matches for my VMs?
What are the resource tradeoffs when there isn’t an exact match?
Where should the instances be located?
What if no match is found?
What happens if my mix of VMs changes?
How does the pricing change if I change location, storage type, or any of the other options?
1. Which EC2 instance types are the right matches for my VM?
Since the configuration of EC2 instance types is fixed, you need to map your VMs to one of the available instance type. The Cost Calculator for AWS Card lets you choose the criteria to do the initial mapping and you have 4 options.
Match by lowest price: This option finds the lowest priced instance from all regions that matches either by vCPU count or vRAM size.
Match by vCPU Count: This option matches by vCPU count and then finds the lowest priced instance from the available choices.
Match by vRAM Size: This option matches by vRAM size and then finds the lowest priced instance from the available choices.
Match vCPU and vRAM Size: This option matches by both vCPU count and vRAM and then finds the lowest priced instance from the available choices.
2. What are the resource tradeoffs when there isn’t an exact match?
Depending on the match criteria that you have chosen, the matched instance may have a resource tradeoff, either vCPU count or vRAM. If there is such a resource tradeoff, you’ll see it as shown below.
3. Where should the instances be located?
Depending on the match criteria and pricing, the Card may match instances in any AWS region. All the matches are overlaid on a map so you can easily find out their geographical distribution. The Card also gives you the option to override the region. This is useful if you want your matches to be found from only one specific region either for performance or compliance reasons.
4. What if no match is found?
For some specific VM configurations, or guest OS types, there may not be a matching EC2 instance type. In this scenarios, the Card will show which VMs could not be matched and why.
5. What happens if your VMs change?
The Card will show cost for each individual VM’s matched instance, a summary for all the VMs that you selected to match against, and a breakout of storage costs. This is automatically updated whenever you modify pricing parameters and whenever the VMs configurations change.
You can also see how many of each kind of EC2 instance type your VMs were matched against.
6. How does the pricing change if I change location, storage type, or any of the other options?
The three main parameters that control instance pricing beyond instance type and location are: term lengths, storage type, and expected utilization.
Term Length: there are two options 1 or 3 years. Amazon offers bigger pricing discount for 3 year term and you can view the discount in the pricing when you select 3 year option.
Storage Type: If the instance will be used primarily for compute and there is no need for persistent storage, then Instance Storage is the cheapest and it is included by default. But for most cases you’ll probably want persistent storage with Amazon EBS (you can find out more about EBS here). The Card will factor your selection into the cost.
Expected Utilization: If you don't need your virtual machine up and running 24/7, you could save lot of money by powering off the virtual machine when not used. Amazon provides three instance utilization categories light, medium and heavy which you can think of in terms of total desired uptime for the instance. The Card will let you change the VM uptime and automatically select the right utilization category that will meet that uptime at the lowest price.
The convenience and benefit of using a public cloud service such as Amazon Web Service is undeniable. However it is non-trivial to figure out how your systems map to Amazon EC2 instance types and how much it would cost to run your VMs in AWS. With CloudPhysics Cost Calculator for AWS you can get a quick estimate for any part of your existing virtualized infrastructure and easily play around with different pricing parameters and options. We hope this makes your decision a bit easier.
That's the CloudPhysics way: taking the guess work out of virtualization management.
P.S. If you are an existing CloudPhysics user you could login today and start using these cards immediately. If you are new, here is the link again to create an account with CloudPhysics and you can get started in 5 minutes!
Krishna Raj Raja