Editor’s note: In honor of Mother’s Day, we look at how 1-800-FLOWERS.COM, Inc. migrated to Google Cloud as part of its digital transformation to quickly deploy seamless and convenient customer experiences across multiple brands on Mother’s Day and every day. 

As a leading provider of gifts designed to help customers express, connect, and celebrate,  1-800-FLOWERS.COM, Inc. has embraced cloud technologies to grow and transform its business through constant innovation. As part of our digital transformation, we recently completed the migration of our ecommerce platform and other services to Google Cloud. We’ve transitioned from a monolithic to a microservices platform, moved many workloads from our on-premises data centers to our Google Cloud environment, and scaled both horizontally and vertically. 

Since our migration, we’ve developed efficient processes to launch new brands, improved the customer experience across all brands, and seen significantly increased site traffic. 

Nurturing a more delightful customer journey

Customer delight is at the core of everything we do. Whether it be with a flower bouquet, a sweet treat, or a personalized keepsake, our mission is to deliver smiles. With the rise of the COVID-19 pandemic, we’ve all been challenged to find unique and safe ways to continue honoring the special connections in our lives and celebrating occasions with loved ones. Our customers have adapted by doing things such as sending gifts to isolated loved ones, sharing the same meal together virtually, or using video to engage with others through group activities like flower arranging and building charcuterie boards. 

The customer experience is a top priority for us, and we constantly look for innovative ways to enhance the customer journey across our ecommerce platform of more than a dozen brands. As a result, we’ve continued to see a rise in demand as customers enjoy the ease and convenience of our site and discover our full family of brands. 

Migrating to a cloud-first mindset

As we’ve continued to innovate and iterate on the customer experience, we knew we wanted to evolve our platform. We wanted to shift to a microservices platform, which would allow our team the opportunity to release updates to our site more often and set up the right continuous integration/continuous deployment (CI/CD) practices. 

Working with Google Cloud, we were able to move our ecommerce platform to the cloud and standardize our site and brand deployment by building one release that could then be repeated across all of our brands. We built everything in a modular fashion, including microservices and      code libraries, so that sites could be easily constructed and replicated for each brand. And because of this, we were able to launch Shari’s Berries extremely quickly after we acquired the brand in 2019.   

Moving our platform to a completely homegrown solution of microservices was a daunting task. But our team handled it beautifully through load-testing, stress-testing, and building new monitoring tools. And with Google Cloud supporting us all along the way, managing the migration process was simple and easy from start to finish. 

Arranging a better bouquet of services

Currently, we’ve migrated every customer-facing touchpoint for all of our brands to Google Cloud—whether it’s on the web or mobile, our AI bots, or our chat interfaces. 

• We run on Google Kubernetes Engine and Istio.
• We have nearly 200 microservices built to help power our entire ecommerce stack across several cloud services running on Google Cloud. 
• We’re utilizing BigQuery for our offline intelligence.

Results are coming up roses

Our new stack on Google Cloud has benefits for both us and our customers. We moved from a session-based to a token-based system, which provides enhanced security as well as a consistent, convenient experience across all our brands. Using service workers and a single-page app, we are able to download all the relevant site content to the browser in under two seconds to create an instant-click experience for each and every customer. We also use Google Analytics to measure our user interactions and provide personalized results to each customer. Our hope is that with this new system, we can learn from customer behavior to offer gift givers a more personalized shopping experience during each visit. 

The benefits of our new tech stack have not only helped us enhance the solutions we offer to customers today, they’ve also enabled us to offer new ones at lightning speed. With our legacy system, we used to release new code once a week or once a month. Now, even during our peak periods, we’re able to release 10 to 15 times a day and can deploy and pivot quickly to create new microservices and microsites on the fly—often without having to touch any code. 

Efficiencies abound     

The benefits of moving our platform to Google Cloud have extended to our internal teams as well. Before the migration, we had only two environments for developing and testing, which made it time-consuming to test updates before they went into production. Now with Google Cloud, we have several different journey teams—which are made up of developers, product owners, and technical owners—all working in several different environments, solving problems, and creating new solutions together. 

Everyone is now empowered to be self-sufficient, developing and releasing microservices on their own when they’re ready. This has given our developers more time to take part in continued development and learning opportunities. For example, we offer lunch-and-learn sessions as well as other resources for everyone to take advantage of so they can continue to learn and refine their skills.

Planting the seeds for future growth

As we look to the future and think about how we help our customers express, connect, and celebrate, we’ll continue to collaborate across teams to deliver solutions that spread smiles. Specifically, we’re exploring additional use of AI to help us better serve our customers across all our brands. 

We’ve enjoyed the ongoing support we’ve received from the Google Cloud team as they help us build new solutions and design a road map for the future. Their support has helped the 1-800-FLOWERS.COM, Inc. team to realize the power of the cloud and bring the very best experience to our customers.

Learn more about 1-800-FLOWERS.COM, Inc., or check out our recent blog about cloud migration for the real world.

Weather forecasting and climate modeling are two of the world’s most computationally complex and demanding tasks. Further, they’re extremely time-sensitive and in high demand — everyone from weekend travelers to large-scale industrial farming operators wants up-to-date weather predictions. To provide timely and meaningful predictions, weather forecasters usually rely on high performance computing (HPC) clusters hosted in an on-premises data center. These on-prem HPC systems require significant capital investment and have high long-term operational costs. They consume a lot of electricity, have largely fixed configurations, and the underlying computer hardware is replaced infrequently.

Using the cloud instead offers increased flexibility, constantly refreshed hardware, high reliability, geo-distributed compute and networking, and a “pay for what you use” pricing model. Ultimately, cloud computing allows forecasters and climate modelers to provide timely and accurate results on a flexible platform using the latest hardware and software systems, in a cost effective manner. This is a big shift compared with traditional approaches to weather forecasting, and can appear challenging. To help, weather forecasters can now run the Weather Research and Forecasting (WRF) modeling system easily on Google Cloud using the new WRF VM image from Fluid Numerics, and achieve the performance of an on-premises supercomputer for a fraction of the price. With this solution, weather forecasters can get a WRF simulation up and running on Google Cloud in less than an hour!

A closer look at WRF
Weather Research and Forecasting (WRF) is a popular open-source numerical weather prediction modeling system used by both researchers and operational organizations. While WRF is primarily used for weather and climate simulation, teams have extended it to support interactions with chemistry, forest fire modeling, and other use cases. WRF development began in the late 1990s through a collaboration between the National Center for Atmospheric Research (NCAR), National Oceanic and Atmospheric Administration (NOAA), U.S. Air Force, Naval Research Laboratory, University of Oklahoma, and the Federal Aviation Administration. The WRF community comprises more than 48,000 users spanning over 160 countries, with the shared goal of supporting atmospheric research and operational forecasting.

The Google Cloud WRF image is built using Google’s MPI best practices for HPC, with the exception that hyperthreading is not disabled by default, and is easily integrated with other HPC solutions on Google Cloud, including SchedMD’s Slurm-GCP. Normally, installing WRF and its dependencies is a time consuming process. With these new WRF VM images, deploying a scalable HPC cluster with WRF v4.2 pre-installed is quick and easy with our Codelab. OpenMPI 4.0.2 was used throughout this work. Google has had good success with Intel MPI, and we intend to study whether further performance gains can be achieved in this context.

Optimizing WRF

Determining the optimal architecture and build settings for performance and cost was a key part of the process in developing the WRF images. We evaluated how to select the ideal compiler, right CPU platform, and the best file system for handling file IO, so you don’t have to. As a test case for assessing performance, we used the CONUS 2.5km benchmark.

Below, the CONUS 2.5km runtime and cost figure shows the run time required for simulating WRF over a two-hour forecast using 480 MPI ranks (a way of numbering processes) for different machine types available on Google Cloud. For each machine type, we’re showing the lowest measured run time from a suite of tests that varied compiler, compiler optimizations, and task affinity.

We found that compute-optimized c2 instances provided the shortest run time. The Slurm job scheduler allows you to map the MPI tasks to compute hardware using task affinity flags. When optimizing the runtime and cost for each machine type, we compared using srun –map-by core –bind-to core to launch WRF, which maps each MPI process to a physical core (two vCPU per MPI rank), and srun –map-by thread –bind-to thread, which maps each MPI process to a single vCPU. Mapping by core and binding MPI ranks to cores is akin to disabling hyperthreading.

The ideal simulation cost and runtime for CONUS 2.5km for each platform is found when each MPI rank is subscribed to each vCPU. When binding to vCPUs, half as many compute resources are needed when compared to binding to physical cores lowering the per-second cost for the simulation. For CONUS 2.5km, we also found that although mapping MPI ranks to cores results in reduced runtime for the same number of MPI ranks, the performance gains are not significant enough to outweigh the cost savings. For this reason, the WRF-GCP solution does not disable hyperthreading by default.

Runtime and simulation cost can be further reduced by selecting an ideal compiler: the figure below (CONUS 2.5km Compiler Comparisons) shows the simulation runtime for the WRF CONUS 2.5km benchmark on eight c2-standard-60 instances, using GCC 10.30, GCC 11.2.0 and the Intel® OneAPI® compilers (v2021.2.0). In all cases, WRF is built using level 3 compiler optimizations and Cascade Lake target architecture flags. By compiling WRF with the Intel® OneAPI® compilers, the WRF simulation runs about 47% faster than the GCC builds, and at about 68% of the cost, on the same hardware. We’ve used OpenMPI 4.0.2 with each of the compilers as the MPI implementation in this work. With other applications, Google has seen good performance with Intel MPI 2018, and we intend to investigate performance comparisons with this and other MPI implementations.

File IO in WRF can become a significant bottleneck as the number of MPI ranks increases. Obtaining the optimal file IO performance requires using parallel file IO in WRF and leveraging a parallel file system such as Lustre.

Below, we show the speedup in file IO activities relative to serial IO on an NFS file system. For this example, we are running the CONUS 2.5km benchmark on c2-standard-60 instances with 960 MPI ranks. By changing WRF’s file IO strategy to parallel IO, we accelerate file IO time by a factor of 60.

We further speed up IO and reduce simulation costs by using a Lustre parallel file system deployed from open-source Lustre Terraform infrastructure-as-code from Fluid Numerics. Lustre is also available with support from DDN’s EXAScaler solution in the Google Cloud Marketplace. In this case, we use four n2-standard-16 instances for the Lustre Object Storage Server (OSS) instances, each with 3TB of Local SSD. The Lustre Metadata Server (MDS) is an n2-standard-16 instance with a 1TB PD-SSD disk. After mounting the Lustre file system to the cluster, we set the Lustre stripe count to 4 so that file IO can be distributed across the four OSS instances. By switching to the Lustre file system for IO, we speed up file IO by an additional factor of 193, which is orders of magnitude faster than a single NFS server with serial IO.

Adding compute resources and increasing the number of MPI ranks reduces the simulation run time. Ideally, with perfect linear scaling, doubling the number of MPI ranks would cut the simulation time in half. However, adding MPI ranks also increases communication overhead, which can increase the cost per simulation. The communication overhead is due to the increased amount of communication necessitated by splitting the problem more finely across more machines.

To assess the scalability of WRF for the CONUS 2.5km benchmark, we can execute a series of model forecasts where we successively double the number of MPI ranks. Below, we show two- hour forecasts on the c2-standard-60 instances with the Lustre file system, varying the number of MPI ranks from 480 to 1920. In all of these runs, MPI ranks are bound to vCPUs so that the number of vCPUs dedicated to each simulation increases with the increase in MPI ranks. While many HPC workloads run best with simultaneous multithreading (SMT) disabled, we find the best performance for CONUS 2.5km with SMT enabled. Thus, the number of MPI ranks in our runs equals the total number of vCPUs.

As you can see, the CONUS 2.5km Runtime & Cost Scaling figure shows that the run time (blue bars) decreases as the number of MPI ranks and the amount of compute resources increases, at least up to 1920 ranks. When transitioning from 480 to 960 MPI ranks, the run time drops, yielding a speedup of about 1.8x. Doubling again to 1920 MPI ranks, though, we obtained an additional speedup of just 1.5x. This declining trend in the speedup with increasing MPI ranks is a signature of MPI overhead, which increases with more MPI ranks.

Determining your best fit

Most tightly-coupled MPI applications such as WRF exhibit this kind of scaling behavior, where scaling efficiency decreases with increasing MPI ranks. This makes assessing cost-scaling alongside performance-scaling critical when considering Total Cost of Ownership (TCO). Thankfully, per-second billing on Google Cloud makes this kind of analysis a little bit easier. As shown above, a second doubling of the count from 960 cores to 1920 cores can provide an additional 1.5x speedup, but at a 32% higher cost. In some circumstances, this faster turnaround may be needed and worth the extra cost.

If you want to get started with WRF quickly and experiment with the CONUS 2.5km benchmark, we’ve encapsulated this deployment in Terraform scripts and prepared an accompanying codelab.

You can learn more about Google Cloud’s high performance computing offerings at https://cloud.google.com/hpc, and you can find out more about Google’s partner Fluid Numerics at https://www.fluidnumerics.com.