I’ve been at Axcient for quite some time. Over four years in fact, and one of the things that really convinced me to join Axcient was our engineering department. It is so satisfying having a development team that is able to see a pain point in the industry. Then is able to architect a solution to resolve that pain, and then is able to deliver a product that exceeds the expectations.
Just about a year or so after I joined Axcient we started hearing bits and pieces about this new technology that our engineers were working on. Groundbreaking and pivotal were a couple of words that were used to describe it. They weren’t wrong. As development continued on the product that came to be named Axcient Fusion, the testing numbers from QA looked off, wrong, too good to be true. Initial images of entire VMware hosts machines offsite and fully protected in less than an hour? Hosts that included devices with over a terabyte of data, some with large databases? Incremental backups of those same devices in 3-5 minutes? We conducted additional extensive testing to ensure we weren’t dealing with some “best case” QA lab environment anomaly. We weren’t.
There are two primary numbers that are critical for any Business Continuity and Disaster Recovery (BCDR) solution like Fusion: Recovery Point Objective (RPO) and Recovery Time Objective (RTO). RPO can be roughly translated to mean how quickly you can get a usable backup of a given device. RTO roughly translates to how quickly you can use that data backup. Those first tests we did showed a sub-hour RPO of these devices regardless of data size. What would the RTO be? The second set of tests involved taking the images of the devices we had protected in the first test and virtualize them to see how quickly they could be made available. This set of tests also exceeded our expectations. The full orchestration process we had built allowed the virtual private cloud to be created, all of the devices to be up and running and ready to be accessed in less than one hour, a sub-hour RTO. Wow.
We deployed Fusion into some early adopters production environments and saw the same results we were seeing in the lab, even with some of the participants bandwidth being slow enough to remind me of dialup. OK a T1 isn’t dial-up, but I’m spoiled now. These results weren’t enough for us. We were confident that we had architected Fusion to work at scale. When I say at scale, I mean the same performance regardless of the number of devices at a given site. So internally we built out a couple of hundred plus machine VMware environments and repeated our tests. The results? Sub-hour RPOs and RTOs. Double wow.
Double wow is right, but also uh-oh. Sub-hour RTOs and RPOs in hundred device datacenter environments might be hard to believe. So how do we gain some independent verification of this performance as we are going to market with this product? We turned to ESG.
Working with ESG to Showcase Axcient Fusion
The Enterprise Strategy Group, or as they are more commonly referred to, ESG is an IT analyst, research, validation, and strategy firm. One of the services ESG offers is technical validation. As they explain on their website, ESG Lab: Validation = Credibility. For the validation of Axcient Fusion’s performance results we worked with Senior Lab Analyst Vinny Choinski. Vinny provides independent, hands-on validation and analysis of emerging storage hardware and software products.
The first step in the validation process with ESG was to demo the Fusion product to Vinny. I showed everything from how easy it is to deploy the AVM to how easy it is to recover the entire environment. After the demo, I shared all of the documentation we had collected from the internal tests that I mentioned above. Vinny was then able to independently audit and review those logs to validate the performance results. The next step was setting ESG up with their own instance of Fusion to allow them to do a complete end to end test from deployment to recovery as the documented in their report.
I won’t repeat their report all their findings in this blog. I will provide a couple of the findings that may be spoilers however. ESG’s lab validation showed the following facts:
- The initial data transfer required for protection was 79% less data than the production data set.
- The incremental transfer reduced the amount of data required for protection by 97%.
- The average incremental transfer duration was four minutes and 45 seconds.
- The fastest recovery observed during testing with 95 virtual machines took only 60 minutes.
As the report states, “Through auditing of real-world POC data and a review of extensive Axcient performance testing, ESG Lab validated that Fusion can help customers realize the benefits that the cloud can deliver. Fusion incorporates the data reduction, network resource management, scheduling, and cloud usability features required to deliver business-class cloud resilience services.”
Can I get a triple Wow?