Shamir Optical used Compute Engine to create a cost-effective, highly available SaaS solution for optical lens manufacturers around the world with Google’s global network.
Google Cloud Results
- Improves latency for customers at 200 labs worldwide
- Cuts costs while enhancing speed and availability
- Takes control of VM deployment and scaling, deploying resources in seconds, not days
As many who need vision correction know getting new lenses can take days, if not weeks. That delay is fast becoming a thing of the past, thanks to the services of Shamir Optical. With only an optical prescription, the company can instantly design and automatically manufacture high-quality lenses at labs all over the world. Today, more than 200 laboratories worldwide use Shamir’s client-facing application to send prescriptions to Shamir’s cloud-based calculation engine, which instructs machines to automatically manufacture precision lenses.
“Its software as a service for lens manufacturers,” explains Saeed Amasha, software R&D Manager at Shamir Optical. “Every frame needs a different kind of lens made for it, so our solution makes complex calculations with all kinds of parameters to manufacture the lenses. Every hour, we receive thousands of jobs from more than 200 labs across the globe, from China, the United States, India, and Europe. Everywhere.”
“We offer a number of services for lens manufacturers. Most are compute-intensive and demand huge resources in terms of CPU, RAM, and storage. All require constant connectivity and availability 24/7, from anywhere in the world. Our most critical technical challenge is to be available all of the time.”—Saeed Amasha, Software R&D Manager, Shamir Optical
For the Israel-based company, delivering rapid responses and high availability for clients worldwide is vital to its business. That’s why Shamir Optical chose to base its solution on Google Cloud Platform (GCP) for a cost-effective global network that the company could rely on.
“We offer a number of services for lens manufacturers,” says Saeed. “Most are compute-intensive and demand huge resources in terms of CPU, RAM, and storage. All require constant connectivity and availability 24/7, from anywhere in the world. Our most critical technical challenge is to be available all of the time.”
A cost-effective, highly available global network
For companies offering Software as a Service (SaaS) solutions, high availability and connectivity are vital, and require a responsive, scalable infrastructure to provide low latency for users. For Shamir Optical, that challenge is magnified by a globally distributed client base looking to access the solution from 200 labs around the world.
“Shamir has been pioneering cloud-based solutions in this industry ever since we migrated from on-premises servers,” says Saeed. “Unfortunately, although the infrastructure we set up with our previous cloud provider delivered the speed, efficiency, and support we demand, that performance came at a very high price. We conducted a market review and received quotes from the biggest industry providers. Google Cloud Platform really stood out, both for its attractive pricing and from the perspective of software development.”
“Google’s global network of data centers has been a real advantage right from the start. We need to provide very short response times for customers at labs all around the world, and since migrating our systems to GCP, we’ve seen major performance improvements in latency and responsiveness.”—Saeed Amasha, Software R&D Manager, Shamir Optical
Shamir enlisted the support of the cloud experts at Sela Group for the migration. “One of our main concerns was that we wanted close support, and Sela’s experts work with us on a daily basis, handling all of our requests and finding solutions for our technical challenges. They’re an outstanding team and they made the migration very easy. Only a few months after building the first test environment we had already moved to production, and we were able to go live with GCP in almost half the time we had estimated.”
By migrating its products to Compute Engine, Shamir has taken control of the management of its infrastructure. Now the company can create new VMs in seconds instead of days, increasing memory and storage through the GCP console with a few clicks, without being dependent on email communication with an external company. Shamir is making further use of that ease of management in its development process, deploying VMs, scaling up dynamically, and improving overall output.
“Google’s global network of data centers has been a real advantage right from the start,” adds Saeed. “We need to provide very short response times for customers at labs all around the world, and since migrating our systems to GCP, we’ve seen major performance improvements in latency and responsiveness.”
A first in Google Cloud Platform migration
Shamir’s SQL Server Always On cluster presented a key challenge for the migration team. Shamir uses a Listener DNS component to hold all of the SQL node’s IP and redirect requests to the active node in the event of downtime. But, at the time of the migration, that DNS component was not supported in GCP networking. To solve the problem, Sela worked with Google engineers and built a special lab environment to test solutions. With the issues that arose during testing ironed out, the cluster was taken to beta, and then live.
“We were the first in the world to utilize SQL Server Always On technology on Google Cloud Platform,” says Saeed Amasha, Software R&D manager. “We achieved data synchronization between three SQL servers in a cluster of three nodes and for two years now, everything has been working smoothly. We’ve had zero downtime.”
“We receive data ranging from the number of jobs per customer to the most popular lens designs. Applying that to Cloud Machine Learning Engine and BigQuery will improve our marketing and production. Meanwhile, we’re enabling client-side applications to send technical information direct to BigQuery for troubleshooting and BI purposes.”—Saeed Amasha, Software R&D Manager, Shamir Optical
Leveraging data to optimize services
For more than two years, Shamir Optical SaaS solutions have delivered improved performance thanks to low latencies on Google’s global network. And by applying machine learning to its billing data, the company identified where and when it was over allocating resources and adjusted consumption in those areas. That generated efficiencies, while Committed Use Discounts saved an additional 25% on some critical resources. Now Shamir is looking to use more of the data it collects to optimize services with machine learning and analytics.
“We receive data ranging from the number of jobs per customer to the most popular lens designs,” says Saeed. “Applying that to Cloud Machine Learning Engine and BigQuery will improve our marketing and production. Meanwhile, we’re enabling client-side applications to send technical information direct to BigQuery, for troubleshooting and BI purposes. Leveraging data is our next step.”