The broadcasting industry has gone through many evolutions since its inception. From linear over-the-air (OTA) to digital & personalized, to standard to ultra high definition, these evolutions were driven by increased demand from viewers who want more choices. The next evolution is happening now, driven by the emergence in cloud computing in a globally connected world. Broadcasters are understanding that the key for long-term success is embracing technical agility while they innovate their business models. Google Cloud technologies can provide a path for continual transformation, empowering broadcasters a multitude of ways to chart their own growth.

As broadcasters evolve their business models and operations for a digital future, evaluating both financial alongside operational benefits will lead to the best outcome. Legacy and siloed media supply chains restrict the ability to deliver content quickly across multiple consumption platforms. By understanding how cloud capabilities can provide cost savings, allow for more efficiency and scale, and open new revenue streams, broadcasters can harness flexible cloud technologies while achieving cost savings and increasing revenue. 

Media workflows in the cloud

Over the last few years we have seen tremendous growth from media companies migrating their supply chains to the cloud. Today, there exists a whole ecosystem of media technologies that are built to take advantage of the cloud. “Does it work on the cloud?” is no longer driving the conversation. Rather, media companies now want to understand how Cloud can integrate with their business and drive better business outcomes. 

Over the last years we have partnered with leading media companies including Grupo Globo, TelevisaUnivision and others to not only migrate their content supply chain to the cloud, but also leverage cloud capabilities to innovate their services to:

1. Increase and streamline content production
2. Distribute personalized content at planet scale
3. Forge deep relationships with their audiences 
4. Identify new monetization opportunities 

Impact of Cloud on Performance & Financials

M&E companies need to be able to provide more content at a quicker pace, with experiences that are seamless and exciting to viewers to retain their attention and dollars. Moving legacy systems and processes to the cloud is an organization-wide commitment, and the journey can pay off financially, while providing M&E companies valuable industry capabilities. With Google Cloud business value engagement framework, we partner to identify where there are opportunities in cost, output, and impact that IT can have. 

Below are some examples of how we have worked with our customers to map organization optimizations  to business drivers

Working Together – How can Google help

Our focus with customers is to help identify and understand the challenges that Media & Entertainment companies have in moving to the cloud, and coming up with the plan and solutions that Google can do to overcome them. Together we commit to understanding your business, both where you are right now in your IT capabilities as well as the progress you want to make to continue providing the best digital capabilities to clients and employees.

As broadcasters move more processes and solutions to the cloud, the exponential effect of harnessing data and AI power will provide incremental business value across all lines of business. Combined, these impacts to a broadcaster allow both operational excellence while optimizing costs as they continue to expand offerings to customers and regions around the world. 

We recognize that every media company’s journey is different and so are expected business outcomes. Google Cloud works closely with customers – partnering every step of the way – to align technology, the media industry, and business outcomes. 

Operating for 23 years, Rustomjee has carved a niche for itself in the ever-growing real estate sector. Rustomjee’s portfolio includes 14.32 million square feet of completed projects; 12 million square feet of ongoing development; and another 28 million square feet of planned development. These projects span the best locations of the Mumbai Metropolitan Region. Rustomjee adds value to the lives of its homeowners through its core business, its corporate social responsibility initiatives, and its philanthropy. The business strives to ensure that every blueprint includes child-friendly spaces for parks, playgrounds, and learning rooms, encouraging families to spend quality time with each other.

In the 17 years Rustomjee’s Corporate Head of Information Systems, V M Samir, has worked with the business, it has grown from 100 employees to more than 800 employees – primarily professionals such as engineers, architects, lawyers, accountants, and regulatory consultants. The remainder work in marketing, sales, administration, security, human resources, and other associated areas.

Historically, Samir says, the real estate industry globally has lagged in embracing new workplace technologies, making building business cases, securing sponsorship, completing implementations, and encouraging adoption a difficult task.

G Suite powers cloud journey

Rustomjee started operations running an on-premises email service. However, the business wanted to upgrade its anti-spam capabilities and, in 2007, turned to G Suite. “There was no practical way I could build an anti-spam engine within that service that could match the power of the G Suite anti-spam engine and the intelligence held within its databases,” says Samir. “Our second key reason for moving was the lower cost of G Suite relative to an on-premises service that required us to spend on compute, backup, storage, and administration resources. Running G Suite would also help ensure users could still access their emails if their machines experienced an outage.”

Finally, G Suite enabled Rustomjee to access and compose emails from any location – a luxury for a real estate organization whose workers often attended construction projects with poor connectivity. “Finally, G Suite was the only service in those days that integrated calendar, meeting, and storage repositories through single sign-on. Google was so far ahead of the curve at that time and we saw the potential of G Suite to transform our communications and ultimately our business.”

“We did not have to shut down the business operations during the migration because Google Cloud Platform complemented every idea we had. And when my business users came to work on the Monday morning [after the final migration], everything was stable and the performance had improved. We told them our infrastructure had changed and they should start thanking Google Cloud!”

—V M Samir, Corporate Head, Information Systems, Rustomjee

The business started its cloud and IT modernization journey by decommissioning its on-premises email servers and using G Suite for Business. It began testing in March 2007 and went live with all production email services for 500 users in June 2008.

The intuitive nature and ease of use of G Suite made the transition seamless. “I did not have to undertake a large-scale change management exercise,” says Samir. “Users bought into the program and acquired the necessary knowledge quickly, meaning our adoption rate was extremely fast.”

With G Suite, email became the new norm to complete a range of tasks at Rustomjee. “We started exchanging CAD drawings, videos, high-resolution images, and other large files with our consultants,” says Samir. “These files had been difficult to store in on-premises environments.”

Rustomjee has improved collaboration and performance with G Suite, primarily due to four services. Gmail enables workers to communicate seamlessly externally and with each other, while Calendar enables them to set up and synchronize meetings. These meetings can be conducted through Hangouts Meet. “If a group of people internally need to discuss a work order or contract, they can coordinate calendars, sit at different locations within our organization, and collaborate using audio and video on a common service,” says Samir. “They can also work from and update a single document in Drive.”

With email and other collaboration applications running smoothly, Rustomjee saw an opportunity to enhance its technology infrastructure. The business had started operations running workplace applications on servers in small air-conditioned rooms. These servers ran databases and applications that sent data across a network to endpoints including desktops and laptops.

Expansion and the changing demands of workplace technology prompted Rustomjee to upgrade its capabilities, and the business commissioned a data center. However, Rustomjee’s continuously fast-growing compute and storage requirements, as well as the need to access new technologies to innovate and compete, quickly strained its technology model. “We required eight weeks each time we needed to add new compute and storage to our environment,” says Samir. “In addition, the heavy investments in our captive data center meant we were unable to easily leverage new technologies and decommission old technologies.”

Public cloud supports growth

Rustomjee reviewed its options and decided public cloud services could best meet its ongoing needs. “The rising cost of maintaining old hardware would force us to refresh data center technologies every five to seven years,” says Samir. “More broadly, a hardware-defined data center could not adapt quickly in the cloud-computing world, making it difficult to align compute with growth.”

The business started by moving its corporate website and microsites to a public cloud service to accommodate increased traffic delivered from a change in business strategy. “We saw an opportunity to shift our marketing and advertising from print advertising to digital platforms,” says Samir. “We wanted to be available to buyers from the point at which they start looking at properties online.”

However, Rustomjee’s digital marketing teams found it difficult to scale instances in line with demand, and work with complex user administration screens and control panels.

“With Google Cloud Platform, we’ve achieved a considerable reduction in timelines and simplified technology management and administration. All of our business users are delighted.”

—V M Samir, Corporate Head, Information Systems, Rustomjee

Google Cloud Platform presented an opportunity for Rustomjee to run its websites and microsites in a reliable, scalable, and responsive infrastructure. “Our evaluation confirmed Google Cloud Platform could support our growing demand for compute and storage,” says Samir. ‘In addition, because we undertake projects that are geographically distributed, networking – not only within the data center, but that could be accessed from any location – was very important. Only Google Cloud Platform had its own networking infrastructure.”

“More broadly, in a cloud environment, we could select any database we needed and start consuming operating systems as a service,” he adds. “In addition, we did not have to invest capital in licenses and hardware and we could always resize network bandwidth. Further, we could take advantage of pay-as-we-use cloud services, and link this to business growth.”

Google Cloud Platform also allowed Rustomjee to reduce the size of the teams needed to manage and operate its infrastructure. “When you have infrastructure running in the data center and in disaster recovery locations, you need to have a large pool of resources working around the clock to ensure constant uptime, sound backup, and good replication measures in place,” says Samir. “With the cloud, everything is simplified and we do not have to consider issues like how many hard drives have failed in on a particular morning.”

The business started testing and created its first virtual machine instance to Google Cloud Platform in April 2017. Because the Mumbai data center was not yet operating Rustomjee initially hosted its instances in the Google Cloud Platform Singapore Region. This initial move enabled the business to reduce the number of required virtual machine instances from six to one. After a year of running the website on Google Cloud Platform, the business found the simplicity of scaling compute resources meant its digital marketing teams were running more campaigns and servicing more leads, while compute spend had fallen by 56 percent. Furthermore, Rustomjee was not experiencing latency that impacted the user experience.

The opening of the Google Cloud Platform Region in Mumbai in late 2017 gave Rustomjee an opportunity to use compute, storage, and networking services located domestically. The business decided to go all-in on Google Cloud Platform and migrated a range of applications to the service.

These included an SAP enterprise resource planning system, running initially on Oracle but moved to the SAP HANA in-memory computing platform, with the assistance of advisory consultants from a leading firm. This system is integral to Rustomjee’s successful operations, meaning it has to be highly available and responsive. “All our financials are captured and stored in this system, as well as projects, materials management, order management, financial systems, and sales ordering systems, both procurement and supply-side,” says Samir. “In short, we cannot live without SAP ERP!”

Rustomjee turned to SAP specialist partner InfraBeat Technologies, a leading network implementation provider, and Google engineers to complete the migration successfully. “We were all in this together and collaborated closely to deliver the project successfully,” says Samir. “We were very impressed by the expertise and skills of everyone involved.”

A nine-day migration

With assistance from Google Cloud and the partners, Rustomjee moved all its SAP workloads, from sandbox to development, development to quality, and quality to production, in just nine calendar days, without impacting the business. “We did not have to shut down the business operations during the migration because Google Cloud Platform complemented every idea we had,” says Samir. “And when my business users came to work on the Monday morning [after the final migration], everything was stable and performance had improved. We told them our infrastructure had changed and they should start thanking Google Cloud!”

Moving to Google Cloud Platform enabled the business to complete certain customer invoicing workloads in just two hours, down from 13 hours previously. Backups of the SAP enterprise resource planning system that had taken up to six hours, were now being completed in six minutes.

“We have now fully embraced a cloud-first approach and have moved 100 percent of our workloads to the cloud. We depend totally on Google Cloud Platform to run our business.”

—V M Samir, Corporate Head, Information Systems, Rustomjee

Eight weeks down to 20 minutes

Rustomjee has also cut the eight-week cycles needed to set up the infrastructure and applications for a new real estate project, and integrate it into the SAP system, to just 20 minutes. “With Google Cloud Platform, we’ve achieved a considerable reduction in timelines and simplified management and administration,” says Samir. “All of our business users are delighted.”

With its website and enterprise resource planning system running successfully on Google Cloud Platform, the business decided to move its virtual application delivery environment into the service. “We decided to move all 800 of our employees into the cloud, so they could access applications through any endpoint, be it a mobile phone, tablet, thin client, desktop, or laptop,” says Samir.

The business decided to replace an existing application virtualization product with public images available on Compute Engine, with testing of Android, iOS and Windows operating systems across a range of devices proving highly successful. “Running the remote desktop service in Compute Engine and using public images enabled us to operate a leaner, simplified desktop as a service environment,” says Samir. “We have been able to deploy business function-specific virtual machine instances on the cloud and compartmentalize data to the business functions that need them. The business completed the move – including 8.1TB of data – in just six weeks. “The only workloads we did not move during that period were intensive workloads for visualization,” says Samir.

However, the four teams that did not use the remote desktop as a service environment – and that created visualization-intensive workloads – were working in traditional ways, compromising productivity and increasing risk. “They used to be given a desk in their respective offices and their workstations loaded with the applications,” says Samir. “There was no way of taking hourly backups of these devices and there was no recovery mechanism for the applications, because they were not in the data center.

“The other challenge was what happens to the data, because we operate out of multiple locations. The users in these four teams always had to come back to pre-assigned desks and continue working from there. This was counterproductive.”

The business subsequently implemented virtual machine instances with GPU access for visualization-intensive workloads. “We started evaluating Nvidia T4 (Tesla) GPUs in the Mumbai Region on 13 February and went live on 11 March in Mumbai,” says Samir. “By moving these 40 team members to Google Cloud Platform, we have enabled them to work with visualization workloads from any location, improving productivity and flexibility.”

Focus on core business

With Google Cloud Platform, Rustomjee can focus on running real estate projects and using technology to enable them. “I don’t have to refresh my technology every five to seven years, and I can leverage new technologies as they come on stream,” says Samir.

Reliable application access

With the flexibility and scalability of Google Cloud Platform, Rustomjee is now well-positioned to support further growth and operate in accordance with the time-limited nature of the real estate industry in India. “If we have promised to hand over the keys to a customer by a certain deadline and we fail to do so, we face a penalty of 10 percent of the cost of the project,” explains Samir. “The Indian Government introduced this regime in 2017 and business and technology have to align with its requirements.

“Google Cloud Platform is the right service to move us forward and enable us to overcome these challenges,” he adds. “We have now fully embraced a cloud-first approach and have moved 100 percent of our workloads to the cloud. We depend totally on Google Cloud Platform to run our business.”

We recently launched Vertex AI to help you move machine learning (ML) from experimentation into production faster and manage your models with confidence—speeding up your ability to improve outcomes at your organization.

But we know many of you are just getting started with ML and there’s a lot to learn! In tandem with building the Vertex AI platform, our teams are dropping as much best practices content as we can to help you come up to speed. Plus, we have a dedicated event on June 10th, Applied ML Summit, with sessions on how to apply ML technology in your projects, as well as grow your skills in this field. 

In the meantime, we couldn’t resist a quick lesson on hyperparameter tuning, because (a) it’s incredibly cool (b) you will impress your coworkers (c) Google Cloud has some unique battle tested tech in this area and (d) you will save time by getting better ML models into production faster. Vertex Vizier, on average, finds optimal parameters for complex functions in over 80% fewer trials than traditional methods. 

So it’s incredibly cool, but what is it?

While machine learning models automatically learn from data, they still require user-defined knobs which guide the learning process. These knobs, commonly known as hyperparameters, control, for example, the tradeoff between training accuracy and generalizability.  Examples of hyperparameters are the optimizer being used, its learning rate, regularization parameters, the number of hidden layers in a DNN, and their sizes.

Setting hyperparameters to their optimal values for a given dataset can make a huge difference in model quality. Typically, optimal hyperparameter values are found via grid searching a small number of combinations, or tedious manual experimentation. Hyperparameter tuning automates this work for you by searching for the best configuration of hyperparameters for optimal model performance. 

Vertex Vizier enables automated hyperparameter tuning in several ways:

1. “Traditional” hyperparameter tuning: by this we mean finding the optimal value of hyperparameters by measuring a single objective metric which is the output of an ML model.  For example, Vizier selects the number of hidden layers and their sizes, an optimizer and its learning rate, with the goal of maximizing model accuracy.
2. When hyperparameters are evaluated, models are trained and evaluated on splits of the data set. If evaluation metrics are streamed to Vizier (e.g. as a function of epoch) as the model is trained, Vizier’s early stopping algorithms can predict the final objective value, and recommend which unpromising trials should be early stopped. This conserves compute resources and speeds up convergence.
3. Oftentimes, models are tuned sequentially on different data sets. Vizier’s built in transfer learning learns priors from previous hyperparameter tuning studies, and leverages them to converge faster on subsequent hyperparameter tuning studies.
4. AutoML is a variant of #1, where Vertex Vizier performs both model selection, and also tunes architectures/non-architecture modifying hyperparameters. AutoML usually requires more code on top of Vertex Vizier (to ingest data etc), but Vizier is in most cases the “engine” behind the process. AutoML is implemented by defining a tree like (DAG) search space, rather than a “flat” search space (like in #1). Note that you can use DAG search spaces for any other purpose where searching over a hierarchical space makes sense.
5. There are times when you may wish to optimize more than one metric. For example, we would like to optimize model accuracy, while minimizing model latency. Vizier can find the Pareto frontier, which presents tradeoffs for multiple metrics, allowing users to choose the appropriate tradeoff. Simple example: I want to make a more accurate model, but would like to minimize serving latency. I do not know ahead of time what’s the tradeoff between the two metrics. Vizier can be used to explore and plot a tradeoff curve, so users can select on the most appropriate one. For example, “a latency decrease of 200ms will only decrease accuracy by 0.5%”

Google Vizier is all yours with Vertex AI

Google published the Vizier research paper in 2017, sharing our work and use cases for black-box optimization—i.e. The process of finding the best settings for a bunch of parameters or knobs when you can’t peer inside a system to see how well the knobs are working. The paper discusses our requirements, infrastructure design, underlying algorithms, and advanced features such as transfer learning that the service provides. Vizier has been essential to our progress with machine learning at Google, which is why we are so excited to make it available to you on Vertex AI.

Vizier has already tuned millions of ML models at Google, and its algorithms are continuously improved for faster convergence and handling of real-life edge cases. Vertex Vizier’s models are very well calibrated and are self-tuning (they adapt to user data), and offer unique power features, such as hierarchical search spaces and multi-objective optimization. We believe Vertex Vizier’s set of features is a unique capability to Google Cloud, and look forward to optimizing the quality of your models by automatically tuning hyperparameters for you.

To learn more about Vertex Vizier, check out these docs and if you are interested in what’s coming in machine learning over the next five years, tune in to our Applied ML Summit on June 10th, or watch the sessions on demand in your own time.