We’ve optimized migrations to public clouds to ensure cost-effective, scalable solutions.
As predicted by Forbes, 83% of enterprise workloads will be in “the cloud” by year’s end 2020, comprised of 20% private cloud, 22% hybrid cloud, and 41% public cloud. In forecasting and preparing for this critical transition, GeorgeJon has tested and optimized migrations of eDiscovery applications to public clouds to ensure a cost-effective, scalable, and performant solution that meets industry demands.
As the leading eDiscovery solutions engineering provider, we embraced the challenge of benchmarking leading applications, placed in cloud environments, against our 15 years of proven best practices in on-prem Kits. The results were a bit surprising, to say the least, with our major discovery being that forklift migrations (copying an entire platform and dropping it into a cloud environment), appealing due to low planning and testing requirements, ultimately result in costly, issue-ridden, painful experiences.
Data Processing Tests
The first major issue we found occurred during benchmarking of data processing speeds. We tested a pair of the industry’s leading applications and found that “Premium” storage was problematic. Running benchmarks from our standard on-premises Kit to a cloud instance with identical CPU and memory specs, running on the cloud provider’s premium SSD, one app experienced a decrease of 31% in throughput per hour on ingestion-only testing. The other app experienced a 15% decrease in ingestion speed and a 45% increase in the time required for export operations. To state the obvious, this meaningful performance lag will have a negative impact on productivity and ROI.
In testing other application functions, we continually found patterns of performance degradation in imaging, exporting, and indexing, with 11%-29% increases in the time required to complete operations in the cloud. During our testing phase, we also ran a file creation process to mimic heavy data loading and imaging operations, emulating the most time-sensitive and mission-critical discovery processes. Our findings yielded a time increase of 23% compared to a parallel, on-premises solution. Once again, not an ideal situation for efficient workflow.
When it comes to small / standalone projects in remote locations, cloud solutions can be cost-effective. However, for large deployments by hosting service providers, law firms, and enterprise users, architectural changes will be required to make cloud-based environments scalable and affordable. eDiscovery, like many industries, can take advantage of burstable, cost-effective VM’s for most resources. However, eDiscovery requirements stray from industry norms when it comes to the robust storage, heavy SQL, and data processing required on a daily basis, none of which are cost-effective when directly deployed in leading cloud provider environments. Degraded performance from these core functions will make or break your application experience and performance.
The Hybrid Solution
Having seen the underwhelming performance offered by a forklift migration to a cloud environment, our crack engineering team took on the task of building a cost-effective, performant, cloud-based solution that would alleviate the shortcomings of leading cloud platforms. After developing a deep understanding of cloud provider capabilities and the related impact on eDiscovery infrastructure, we designed a forward-thinking solution that meshes proven, on-premises performance with cloud-based scalability and on-demand resources. The result is our revolutionary Kit-as-a-Service (KaaS) platform, uniting the best aspects of both models to deliver a cost-effective, endlessly scalable, secure and reliable solution for clients worldwide!
If you found this information helpful and would like to discuss hybrid on-prem/cloud hosting options that deliver endless flexibility and industry-leading performance, please contact us.