Everything you need to know about the stuff that clouds are made of.
Amid the twists and turns in the current business environment brought about by Covid-19, the shift towards cloud services may be one that is here to stay.
Cloud computing helps businesses circumvent onsite infrastructure costs by hosting databases or software on remote servers instead, which are then accessed through the Internet. Combined with the everything-as-a-service economy (XaaS), working in the cloud has become a hallmark of the virtual age.
One study highlights that spending on cloud services touched $31 billion in the first quarter of 2020, as compared to $23.1 billion in the same timeframe last year, with a consistent growth pattern.
Another report suggests that there has been an increase in demand for cloud services from healthcare, manufacturing, and banking and financial sectors, as well as for remote cloud collaborative tools such as Slack or Zoom.
Ensuring business continuity is one of the key forces behind this demand, Benjamin Hui tells Jumpstart.
Hui is a solutions engineer at Hong Kong-based cloud data services provider NetApp, where he is responsible for building Netapp’s hybrid multicloud solutions to help enterprises clients manage their data.
How can cloud storage be used for backups?
In a place like Hong Kong that is dealing not only with the Covid-19 pandemic but also with a tense geopolitical climate, Hui observes that NetApp’s clients want to ensure business continuity without having to invest in additional hardware. This is where cloud services come in.
Even if something untimely happens to data centers kept on premises, cloud services can help the business continue functioning without having to make the expensive call of rebuilding their on-premises infrastructure or deal with loss of data.
“A lot of our customers invest in the cloud with an incremental approach. They put some stuff in the cloud to a maintain the minimal working level, usually based on budget and space,” he says.
Lockdown measures have also led to a spike in remote workloads for several companies, which may not be sustainable as the world adjusts to a new normal. A shorter term investment in cloud services, which can be availed on a subscription or pay-as-you-go model, is a cost-effective solution, Hui says.
“It doesn’t make sense for a startup to buy a lot of hardware to cater for this surge in demand. And this is also why cloud is very useful right now,” he notes.
In a world under lock down, transitioning to remote working has been easier for younger companies that are ‘born in the cloud’, or are cloud-native.
These companies have been using cloud services from the get-go, well before the global crisis, to keep the cost of building their data centers or servers from eating into their profit margins.
However, more traditional businesses or those only starting to migrate some of their workload to the cloud are at a slight disadvantage, now dealing with the complexities of cloud data management in a rush due to the suddenness of the work-from-home experiment.
Matters of security
The primary concern with cloud solutions is data sovereignty and cyber security. Data is a business asset, and companies need to protect access to it. Cloud platforms are vulnerable to numerous security risks such as data breaches, malicious software, phishing, or identity theft.
This may put sensitive information in the hands of hackers, or crash cloud systems, as happened with the eight-hour distributed denial of service (DDoS) attack on Amazon’s AWS late last year.
Hui notes that policies and data protection methods differ from provider to provider, and it is on the companies to find a way to align this with their own security protocols.
“It is not that the cloud is safe or not safe, it is the party that needs to keep it safe,” he points out.
Apart from training employees on cybersecurity hygiene measures, or using multi-factor authentication for added security, Hui suggests that it can be helpful to maintain backups or utilize multiple platforms simultaneously to ensure continuous availability of data.
For instance, a Hong Kong-based client shifted all their data to a center in Singapore when Hong Kong’s pro-democracy protests got heated up, Hui recalls.
This also means that companies hosting data at multiple offshore locations will need to be cautious with regulatory compliance, which vary from country to country.
Compliance requirements, which are jurisdictionally regulated by government agencies such as the Monetary Authority of Singapore or Hong Kong Monetary Authority, require data flexibility on the company’s front, especially for those that operate in multiple countries, or host their data on servers outside their country of business.
“Different countries, especially in Southeast Asia, have different ideas of cloud and cyber security and compliance. You don’t want to redesign your entire solution when [regulators] say it is okay to go to the cloud [only to take it down later],” Hui explains.
He continues, “That’s why we see a lot of people looking for what we call Data Fabric (a technology that helps companies manage data across multiple databases and systems) or some data management solution across all premises. They want to be flexible, yet they want to be safe.”
Complexities of cloud migration to watch out for
Operational and data complexity is another circumstance that companies will need to navigate carefully when moving to cloud, Hui believes. Businesses operate through teams, and these teams need to work in sync with each other.
“In the on-premises world, the operations team or the IT admin team manages resources, and in the cloud world the development team usually gets more freedom,” he says. “This has increased the difficulty for the operations teams, since they need some kind of way to track the bills. You don’t want to get a million bills from the cloud after a month putting you out of business.”
Hui notes that hiring candidates in a DevOps role, which combines software development and IT operations, can help solve the quandary, leading to “better collaboration of culture between deployment teams and operations teams.”
Data complexity poses another challenge to companies. Companies can host data on private cloud which could be on-premises, or public cloud such as AWS and Microsoft Azure, or a combination of both, called hybrid cloud.
They can also use cloud services from multiple service providers, called multicloud, so that their data is distributed across vendors and not locked in by one.
If that is not enough, companies can combine these two frameworks to employ a hybrid multicloud approach for data management, which is what NetApp specializes in. As each of these layers get added on, access and movement of information gets more complex and expensive.
“There’s a saying that moving to the cloud is easy, but moving out of the cloud is very hard, because clouds, actually, do not charge you for ingress (taking in data), but will charge you an enormous amount for egress (moving the data out from the cloud),” Hui explains.
He notes that a lot of people may not even have the expertise to build such an elaborate system, so even though the idea is popular, supply of cloud architects in the market does not match up.
“It’s like Walmart”
While cloud technology is definitely here to stay, several trends have emerged of late due to rapid cloud storage adoption and digital transformation.
While virtual desktop infrastructure, where desktops are virtualized through onsite technology, is one of the most sought after cloud products right now, a more powerful tool that is driven by advanced technologies is the data lake, according to Hui.
“Data is growing exponentially and people need a scalable solution,” he says. “We’re seeing a lot of new workloads and adoption in artificial intelligence (AI) and 5G. A lot of startups are using Internet of Things (IoT) devices. And they’re looking for cloud-scale data lake technology.”
A data lake, Hui explains, is a place to put all of the enterprise’s data, and use deep learning and AI to analyze it for trends and patterns. Where terabytes were once considered a colossal amount, today’s data is dealt with in petabytes–one thousand million million bytes–and needs solutions powerful enough to process that amount.
Another trend that Hui is excited about is Kubernetes, something that all cloud service providers are scrambling to provide.
Announced by Google in 2014, Kubernetes is an open-source platform that automates the deployment, scaling, and management of applications through what is called ‘containerization’. Simply put, it involves placing applications in stand-alone containers with their own operating systems, making the applications portable.
“In older times, we had physical servers which were divided into multiple smaller servers, called virtual machines. In this new world, we can slice even smaller machines out of virtual machines, called containers,” Hui explains. He adds that Kubernetes is an open-source platform for managing containers–and every cloud storage provider is “selling it like hotcakes.”
Hui adds that while Kubernetes has built a more effective way to move applications, the challenge is to find a way to facilitate the movement of data between clouds as well, something that he has already described as an expensive process.
AI in cloud is another popular development, which Hui believes has moved past the hype and has actual utility. From fraud detection in banks, to AI infrastructure for research on Covid-19, to autonomous vehicles and quantum computing astrophysics, the applications of AI in cloud span entire industries.
However, the tech does not come cheap, since the graphics processing unit required to power AI can be heavy on the pocket, although Hui believes that the costs will come down as more takers for the tech enable economies of scale.
Choosing the right cloud provider for your business
Not every cloud option will be the right fit for your business, and finding the right combination that is cost-effective, secure, and solves business problems requires balanced decision-making.
One of the things to focus on, according to Hui, is the risk of being locked in, as different vendors may have different lock-in periods for their products, services or platforms.
“This affects how many clouds you’re using, whether it’s on-premises (so you can control everything) or multicloud. You don’t want to be controlled by just one provider, and you want to look into other providers when the price surges,” he says.
Another aspect to consider is how complex your operations will be and how much flexibility you want between clouds.
“You need to have a common data plane to make sure that moving data between them is cost effective. You also need to think about how to protect data in a distributed infrastructure. So maybe hot data is kept on premises, and cold data is moved to the external cloud provider which is cheaper,” he suggests.
Ultimately, Hui sees cloud storage as a supermarket and upcoming innovations as products with different applications that need to be assessed on a use-case basis. Much like toilet paper and tissues, “you need to use the right tool for the right job,” he quips.
Header image courtesy of NetApp