HealthCheck for Azure observes network traffic for a two-week period, assembles the data it collects, and then provides an assessment report. The report details a summary of the network infrastructure and traffic, including both physical and virtual network components.
With the report in hand, you can present your customers with sound, data-driven recommendations about moving their infrastructure (or portions of it) to Microsoft Azure. The report has three main sections: Infrastructure Overview, Approach, and Total Cost of Ownership.
The Infrastructure Overview section lets you compare various topologies or designs. It provides data on the following data points:
|Total virtual nodes||Number of network nodes, both physical machines and virtual instances.|
|Total virtual disks||Number of virtual and physical disks.|
|Total bytes per month (GB)||Total network bndwidth consumed, both internal to the data center, and bytes served to the web.|
|Hardware mapping||Total cost of a cloud deployment, assuming a lift-and-shift approach with existing hardware.|
|Workload mapping||Total cost of a cloud deployment, assuming the elimination of over-provisioned instances.|
This section of the Assessment report examines the costs and benefits of the three main approaches to an Azure
migration. It reports on hardware mappings, workload mappings, as well as
whether standard or premium storage is best. It also provides data on the
overall utilization of network devices, including CPU and RAM. Finally, you
receive an explanation of the hardware and workload mappings, as well as the
pay-as-you-go pricing plan.
|Hardware mapping||This approach takes a "lift-and-shift" approach to migrating the existing infrastructure to Azure. This approach is typically the most expensive.|
|Workload mapping||This approach looks at the load on each server and tries to right-size the cloud image in cases where the original image is overprovisioned. This approach normally results in significant cost savings.|
|Pay-as-you-go pricing||The pay-as-you-go plan requires that you pay only for actual usage. |
Total Cost of Ownership
This section of the Assessment report provides total cost of ownership (TCO) estimates on four items:
- Compute: The cost for the CPU and RAM mappings.
- Storage: The cost for physical storage, as well as for observed IOPS requirements.
- Network: The cost for total network usage. Note that "total" usage includes both internal data center traffic as well as data served to the web.
- Total: The sum of all costs.
The report provides a chart that compares three pricing plans reflecting three approaches to hardware mapping:
- Hardware Mapping Pay-as-you-go Pricing Plan. This is the most expensive of the plans, as it does not try to find overprovisioned servers to map to more appropriate servers, and it uses pay-as-you-go pricing.
- Workload Mapping Pay-as-you-go Pricing Plan. The customer should see significant savings here, as many in-house systems are overprovisioned and can be mapped to smaller, more appropriate instances. It uses pay-as-you-go pricing.
- Workload Mapping Pre-Pay 12 months. This option combines the savings of workload mapping, with a pre-pay discount offered by Microsoft Azure.
Following the TCO report section, each of the three mapping plans is broken out with additional columns:
- Data Center. If the customer grouped their servers into data centers, the data center name is listed. If the data center is "total", this is the aggregate of all reported data centers.
- Virtual Instances. The number of observed Virtual and Physical Machines.
- Estimated Annual Compute Cost. The expense to provision the virtual instances in this mapping (CPU and RAM).
- Virtual Disks. The number of observed physical and virtual disks reported to Cloudamize.
- Virtual Storage. The observed amount of Disk Space needed.
- Estimated Annual Compute Cost. The cost to store the observed data.
- Storage IO Costs. The cost to purchase Premium or Standard storage based on application usage/throughput.
- Network IO. The cost to provision predicted network traffic. Note that this includes intra-data center traffic.