BRIAN FAWCETT (BF): The global silicon chip shortage is complicating matters for organizations around the world, making it difficult for many to test development, support the front end of their business and deliver services. Sadly, it's likely going to get worse before it gets better.
I'm your host Brian Fawcett and this is IT Availability Now, the show that tells stories of business resilience, from the people who keep the digital world available.
On this episode of IT Availability Now, Servaas Verbiest, Lead Cloud Evangelist at Sungard AS, discusses how the chip shortage is influencing business decisions, pushing companies to rethink their entire IT model and accelerating cloud adoption.
Servaas, welcome back to the show.
SERVAAS VERBIEST (SV): Thanks for having me, Brian.
(BF): So Servaas, can you give us a bit of a rundown on what's really happening right now with this chip shortage and how it's impacting businesses?
(SV): So, it's a really interesting dynamic because when things came to a halt during the initial phases of COVID, a lot of things were impacted and one of those was the production of the silicon chips that are prevalent in a lot of technology across a bunch of consumer and business products, right? And as business started to spin up again and the world started to go back to the new normal, the demand for those chips was where it was prior to things shutting down, and the way we structured logistics was such that everything was running in the most efficient fashion, producing only what they needed for a given time.
Well, it had to ramp back up. And because of that latent demand, you now have a scenario where businesses that are looking to expand their infrastructure are vying with auto manufacturers and, you know, cell phone providers for vital components that actually power those pieces of equipment. So you end up having providers vying for this limited commodity, which impacts the timeline to actually gain access to these mission critical components to infrastructure. I'm working with clients right now who are in a position where they're starting to value their position on cloud and the consumption of hardware a little differently, because they're being given lead times of almost 200 days in some circumstances to get access to what they need.
(BF): So, it sounds like this chip shortage is really forcing businesses’ hand to consider moving to the cloud. How are businesses reacting to this and what are their biggest concerns?
(SV): So, the cloud really is just a financial model for the consumption of compute, RAM and storage, and maybe some high-level services that provide, you know, some of the features and functions that an application does. But that said, every use case and every workload has performance requirements, and those performance requirements have different financial models that are aligned to them. So, as you look at these use cases and how applications need to be delivered and what they support, you'll find that there are different financial models that are just going to make more sense.
And if you're in a position where you're looking to grow your business, and you're traditionally familiar with, you know, more of a hardware-based model, or you started with a little bit of cloud and you're starting to begin embracing a cloud consumption model, being forced to fully embrace it at a rapid pace leaves you very exposed. If you make the wrong decisions, you can have performance impacts that have downstream impacts on your end users and your customer base, which can have market ramifications.
And if you don't consider the financial requirements associated to maintaining the right level of performance that you need to meet the expectations of your internal and external customers, you may find your business in a position where it's over leveraged financially, and now you've got to make some hard decisions based on what you’ve designed.
(BF): So, let's dive into that a bit more and let's start with the financial models. Can you walk us through the various models and talk a little bit about the pros and the cons of each of us?
(SV): Okay, let's start with the one that a lot of people are familiar with and what they used to use but are having problems with as a result of logistics and chips shortage. If you buy a physical box, right, and it's a workload that runs consistently 24/7, 365 and its performance requirements don't change, you can recognize the cost associated with that asset and over time, depending on the accounting principles that you follow, and how your organization operates, depreciate it. And you don't have to worry about some of the complexities of capacity planning and meeting other requirements. So, it's going to be a great model for you.
But, you know, when you start to look at leveraging cloud technologies, if you've got a more volatile workload or you have a workload that requires delivery into a unique region where you don’t have a space or you've got a workload that, while it's not volatile, it's going to keep getting bigger, a cloud consumption plan could be fantastic for you. The challenge is, you know, there's a lot of different ways to consume that. You can pay for it as you go, which if you look at the units of measurement associated with that particular cloud service, you just got to be cognizant of all those levers. They have opportunities to reserve components of cloud compute, where you commit to a certain configuration over time, and that affords you a discount on those resources, so you have a lower unit price, but you're pretty much locked into that configuration for as long as that commit has to run its course. So if there's any deviation, you're paying that existing commitment, and you're paying for whatever you use on top of that.
And then there are other components like, you've probably heard of spot instances where you're paying for access to a certain compute model for a short period of time, or even commodities markets that are arising where people were taking these commitments that they have, and then reselling them to others that need them so they don't have to maintain those commitments. And it turns into a really interesting situation where you're almost working through commodity trading with cloud resources, so it can get pretty complicated.
(BF): That's very interesting. What about the performance ramifications of leveraging these cloud technologies? Can you get into those a little bit more?
(SV): So ideally, you want to match the performance requirements with the expectations to what is fiscally responsible for the deployment. You need to perform an analysis to determine what the workload needs to run on at the level that you desire, and the specific speed and performance requirements from a compute, RAM and storage perspective so you're not under or over allocating what you need.
(BF): And is there a way to take a similar approach in the cloud?
(SV): So, if you work with an organization or you have the internal skill set in house, which you know, some organizations do, and you have the right tooling that allows you to track utilization over peak and low periods, and then you have financial assets that understand these various financial models and cloud that can match them up to the performance requirements provided by your technical resources, you can structure a cloud deployment that maximizes your rate of consumption and the costs associated to it. Most organizations don't necessarily have all those skills. There are gaps, whether it's through things like the Great Resignation with people stepping away to move to a different organization because maybe they found a better opportunity, or changes in the marketplace that require businesses to restructure and really reconsider how they operate.
That's why when I bring up a managed service provider like ourselves, we can fill those gaps because we have an understanding of not only what it takes to design these environments, but we can properly display, manage, maintain and continue to optimize over time, their financial structure and performance so that you get the best bang for your buck. And if you could maintain discipline in that process, embracing a cloud model to maintain or accelerate your speed to market and balance it with existing hardware or fiscal deployments where the use case permits it, you're going to be in a great spot even with the limitations around the number of chips that are available.
(BF): That’s excellent Servaas. For organizations that are being forced to quickly accelerate their cloud adoption, what can they do to minimize their costs.
(SV): Okay, so when you talk about cost optimization, right, it's a very popular topic, it comes up quite a bit. It's about being disciplined and having a process that allows you to audit, review and control what you deploy. A lot of organizations get into trouble because they start leveraging cloud technologies and it's something that starts off as a hobby and then becomes valuable to the business and becomes integrated into standard business operations. With that said, you've got a lot of people who have to make sure that it checks a lot of boxes, so you don’t violate security requirements or create any performance impacts or have any data sprawl because of the resources that are being integrated now.
So if you can, from an operational perspective, put governance in place that allows you to review, maintain and control what is deployed, and work closely with someone that understands those cloud financial models to make sure that there's no wasted spend, or if there's an opportunity based on utilization, to embrace something like a reserved instance or, you know, some providers offer reserved instance savings plans, you can, right? And, just to provide a high level example, when you talk about these reserved instances, a good rule of thumb is, if the workload is going to be running greater than 49% of the time, a reserved instance can be a great fit, because at the end of the day, you're going to make out better. When that threshold doesn't get crossed, and you find yourself committed to a reserved instance, you know, that's where you end up coming down on the losing end. And if you do that enough times across the wide area network with a large footprint, it can be very detrimental to the organization.
(BF): Excellent recommendations. There are only so many silicon chips that can be manufactured in a given time, and organizations across multiple industries are vying for access to them. Unfortunately, the global shortage is only adding to the stress. Many companies are turning to cloud technologies to avoid logistical nightmares.
But as cloud adoption accelerates, businesses must be sure they’re making the right strategic decisions and getting maximum value. Performing a workload analysis can help you match performance requirements and expectations to a fiscally responsible deployment, ensuring that your business won’t be hindered by a sudden shift in its IT strategy.
Servaas, thanks for joining the discussion today.
(SV): Thanks for having me, Brian.
(BF): Servaas Verbiest is Lead Cloud Evangelist at Sungard AS.
You can find the show notes for this episode at SungardAS.com/ITAvailabilityNow.
Please subscribe to the show on your podcast platform of choice to get new episodes as soon as they’re available.
IT Availability Now is a production of Sungard Availability Services.
I’m your host, Brian Fawcett, and until next time, stay available.