Blog Viewer

It's Difficult to Ignore the Data Protection Market

By Lori Donaldson posted 06-20-2018 09:33 AM

  

It's Difficult to Ignore the Data Protection Market

Data management enables data mobility. An advantage to data mobility is improved disaster recovery capability. Looking to the public cloud - an obvious temptation is to leverage public cloud as an on-demand disaster recovery site.

 

As administrators, we don’t talk about the economics of Silicon Valley startups often. However, it’s difficult to ignore the data protection market. Companies ranging from Rubrik and Cohesity to Druva Inc., have all had $90 million or more invested in each company over the past few months. It makes you sit up and wonder what’s so special about backup. Rubrik won Best of Show at VMworld 2017. While backup is intriguing the real value of this category is data management features.

 

Modern backup applications all aim to become data management platforms. An added benfit for backup software? The capability to replicate snapshots to the public cloud in increments as small as 15-mins. The result is a recovery point objective (RPO) of as low at 15-mins. Best of all, you don’t have to pay for idle capacity. Your only cost outside of cloud storage comes when enacting a test of failover. On paper, it’s a game-changing capability. In reality, there are a lot of moving parts. This post will examine 3-areas of focus.

 

1 – VMDK format conversion

All the modern-day backup solutions make getting the data to the public cloud reasonably simple. It’s bringing up the environment that proves the challenge. Once you have a VMDK file in AWS, what do you do with it? Unless you’re using VMware Cloud on AWS, then you can’t merely boot a VMDK as an EC2 instance. You must first convert the VMDK.

For smaller environments, manually converting a few VMs to EC2 isn’t too troublesome. However, once you reach a dozen or more instances, the challenge is considerable. Your recovery time objective (RTO) becomes increasingly longer as your environment grows in size. Many modern-day backup products allow for the automation of converting a VM to a public cloud format. Check with your backup vendor to check which public clouds your solution supports.

 

2 – Network

Getting your VMs up in running solves the first problem. Another challenge is networking. Again, if you have a small environment with a flat network, then this isn’t much of a challenge. However, if you have a sophisticated data center with multiple subnets, then you must consider a virtual network design that supports a similar design. That includes any access controls between subnets. Implementing and keeping these cloud access control rules in-sync with your private data center takes consideration. VMware NSX-T might prove helpful.

 

Outside of data center connectivity, here is a list of additional network-centric concerns.

  • How will users access the virtual data center you’ve stood up in the cloud?
  • Will you leverage VDI or VPN?
  • What software or settings need to be distributed to end users to access the public cloud resources.
  • Have you integrated additional SaaS services that require firewall rules to access resources?

 

3 – Continuous operations

Even if you’ve bested these design challenges, an additional challenge is keeping configurations in-sync post-implementation. Data centers are living systems. They change day-by-day and hour-by-hour. What worked in a DR test last week doesn’t work in an enactment next week. A critical lesson I’ve learned over the years – DR operations should mimic production.

 

The public cloud proves the perfect implementation of the software-defined data center (SDDC). You should strive to implement infrastructure as code practices in the private DC. If you’ve automated provisioning in your private data center, you can leverage the same processes to automate recovery in public cloud. If you’ve automated network changes, you can automate network configuration in the public cloud.

 

Disaster recovery is hard no matter the technology used. It’s one of the best examples of people, process, and technology all working together. If you use the same processes in DR as you do in production, then your chances of successfully leveraging public cloud for DR increases.

Keith Townsend, Owner of The CTO Advisor and VMUG Contributor

0 comments
3 views