Category: Cloud Infrastructure

Sustainable Business Needs Sustainable Technology: Can Cloud Modernization Help? | Blog

Estimates suggest that enterprise technology accounts for 3-5% of global power consumption and 1-2% of carbon emissions. Although technology systems are becoming more power efficient, optimizing power consumption is a key priority for enterprises to reduce their carbon footprints and build sustainable businesses. Cloud modernization can play an effective part in this journey if done right.

Current practices are not aligned to sustainable technology

The way systems are designed, built, and run impacts enterprises’ electricity consumption and CO2 emissions. Let’s take a look at the three big segments:

• Architecture: IT workloads, by design, are built for failover and recovery. Though businesses need these backups in case the main systems go down, the duplication results in significant electricity consumption. Most IT systems were built for the “age of deficiency,” wherein underlying infrastructure assets were costly, rationed, and difficult to provision. Every major system has a massive back-up to counter failure events, essentially multiplying electricity consumption.
• Build: Consider that for each large ERP production system there are 6 to 10 non-production systems across development, testing, and staging. Developers, QA, security, and pre-production ended up building their own environments. Yet, whenever systems were built, the entire infrastructure needed to be configured despite the team needing only 10-20% of it. Thus, most of the electricity consumption ended up powering capacity that wasn’t needed at all.
• Run: Operations teams have to make do with what the upstream teams have given them. They can’t take down systems to save power on their own as the systems weren’t designed to work that way. So, the run teams ensure every IT system is up and running. Their KPIs are tied to availability and uptime, meaning that they were incentivized to make systems “over available” even when they weren’t being used. The run teams didn’t – and still don’t – have real-time insights into the operational KPIs of their systems landscape to dynamically decide which systems to shut off to save power consumption.

The role of cloud modernization in building a sustainable technology ecosystem

In February 2020, an article published in the journal Science suggested that, despite digital services from large data center and cloud vendors growing sixfold between 2010 and 2018, energy consumption grew by only 6%. I discussed power consumption as an important element of “Ethical Cloud” in a blog I wrote earlier this year.

Many cynics say that cloud just shifts power consumption from the enterprise to the cloud vendor. There’s a grain of truth to that. But I’m addressing a different aspect of cloud: using cloud services to modernize the technology environment and envision newer practices to create a sustainable technology landscape, regardless of whether the cloud services are vendor-delivered or client-owned.

Cloud 1.0 and 2.0: By now, many architects have used cloud’s run time access of underlying infrastructure, which can definitely address the issues around over-provisioning. Virtual servers on the cloud can be switched on or off as needed, and doing so reduces carbon emission. Moreover, as cloud instances can be provisioned quickly, they are – by design – fault tolerant, so they don’t rely on excessive back-up systems. They can be designed to go down, and their back-up turns on immediately without being forever online. The development, test, and operations teams can provision infrastructure as and when needed. And they can shut it down when their work is completed.

Cloud 3.0: In the next wave of cloud services, with enabling technologies such as containers, functions, and event-driven applications, enterprises can amplify their sustainable technology initiatives. Enterprise architects will design workloads keeping failure as an essential element that needs to be tackled through orchestration of run time cloud resources, instead of relying on traditional failover methods that promote over consumption. They can modernize existing workloads that need “always on” infrastructure and underlying services to an event-driven model. The application code and infrastructure lay idle and come online only when needed. A while back I wrote a blog that talks about how AI can be used to compose an application at run time instead of always being available.

Server virtualization played an important role in reducing power consumption. However, now, by using containers, which are significantly more efficient than virtual machines, enterprises can further reduce their power consumption and carbon emissions. Though cloud sprawl is stretching the operations teams, newer automated monitoring tools are becoming effective in providing a real-time view of the technology landscape. This view helps them optimize asset uptime. They can also build infrastructure code within development to make an application aware of when it can let go of IT assets and kill zombie instances, which enables the operations team to focus on automating and optimizing, instead of managing systems that are always on.

Moreover, because the entire cloud migration process is getting optimized and automated, power consumption is further reduced. Newer cloud-native workloads are being built in the above model. However, enterprises have large legacy technology landscapes that need to move to an on-demand cloud-led model if they are serious about their sustainability initiatives. Though the business case for legacy modernization does consider power consumption, it mostly focuses on movement of workload from on-premises to the cloud. And it doesn’t usually consider architectural changes that can reduce power consumption, even if it’s a client-owned cloud platform.

When considering next-generation cloud services, enterprises should rethink their modernization journeys beyond a data center exit to building a sustainable technology landscape. They should consider leveraging cloud-led enabling technologies to fundamentally change the way their workloads are architected, built, and run. However, enterprises can only think of building a sustainable business through sustainable technology when they’ve adopted cloud modernization as a potent force to reduce power and carbon emission.

This is a complex topic to solve for, but we all have to start somewhere. And there are certainly other options to consider, like greater reliance on renewable energy, reduction in travel, etc. I’d love to hear what you’re doing, whether it’s using cloud modernization to reduce carbon emission, just shifting your emissions to a cloud vendor, or another approach. Please write to me at [email protected].

Cloud Wars Chapter 5: Alibaba, IBM, and Oracle Versus Amazon, Google, and Microsoft. Is There Even a Fight? | Blog

Few companies in the history of the technology industry have aspired to dominate the way public cloud vendors Microsoft Azure, Amazon Web Services, and Google Cloud Platform currently are. I’ve covered the MAGs’ (as they’re collectively known) ugly fight across other blogs on industry cloud, low-code, and market share.

However, enterprises and partners increasingly appear to be demanding more options. That’s not because these three cloud vendors have done something fundamentally wrong or their offerings haven’t kept pace. Rather, it’s because enterprises are becoming more aware of cloud services and their potential impact on their businesses, and because Alibaba, IBM, and Oracle have introduced meaningful offerings that can’t be ignored any longer.

What’s changed?

Our research shows that enterprises have moved only about 20 percent of their workloads to the cloud. They started with simple workloads like web portals, collaboration suites, and virtual machines. After this first phase of their journey to the cloud, they realized that they needed to do a significant amount of preparation to be successful. Many enterprises – some believe more than 90 percent – have repatriated at least one workload from public cloud, which opened enterprise leaders’ eyes to the importance of fit-for-purpose in addition to a generic cloud. So, before they move more complex workloads to the cloud, they want to be absolutely sure they get their architectural choices and cloud partner absolutely right.

Is the market experiencing public cloud fatigue?

When AWS is clocking over US$42 billion in revenue and growing at  about 30 percent, Google Cloud has about US$15 billion in revenue and is growing at over 40 percent, and Azure is growing at over 45 percent, it’s hard to argue that there’s public cloud fatigue. However, some enterprises and service partners believe these vendors are engaging in strongarm tactics to own the keys to the enterprise technology kingdom. In the race to migrate enterprise workloads to their cloud platforms, these vendors are willing to proactively invest millions of dollars – that is, on the condition that the enterprise goes all-in and builds architecture for workloads on their specific cloud platform. Notably, while all these vendors extol multi-cloud strategies in public, their actual commitment is questionable. At the same time, this isn’t much different than any other enterprise technology war where the vendor wants to own the entire pie.

Enter Alibaba, IBM, and Oracle (AIO)

In an earlier blog, I explained that we’re seeing a battle between public cloud providers and software vendors such as Salesforce and SAP. However, this isn’t the end of it. Given enterprises’ increasing appetite for cloud adoption, Alibaba, IBM, and Oracle have meaningfully upped the ante on their offerings. The move to the public cloud space was obvious for IBM and Oracle, as they’re already deeply entrenched in the enterprise technology landscape. While they probably took a lot more time than they should have in building meaningful cloud stories, they’re here now. They’re focused on “industrial grade” workloads that have strategic value for enterprises and on building open source as core to their offering to propagate multi-cloud interoperability. IBM has signed multiple cloud engagements with companies including Schlumberger, Coca Cola European Partners, and Broadridge. Similarly, Oracle has signed with Nissan and Zoom. And Oracle, much like Microsoft, has the added advantage of having offerings in the business applications market. Alibaba, despite its strong focus on China and Southeast Asia, is increasingly perceived as one of the most technologically advanced cloud platforms.

What will happen now, and what should enterprises do?

As enterprises move deeper into their cloud journeys, they must carefully vet and bet on cloud vendors. As infrastructure abstraction rises at a fever pitch with serverless, event-driven applications and functions-as-a-service, it becomes relatively easier to meet the lofty ideals of Service Oriented Architecture to have a fully abstracted underlying infrastructure, which is what a true multi-cloud environment also embodies. The cloud vendors realize that, as they provide more abstracted infrastructure services, they risk being easily replaced with API calls applications can make to other cloud platforms. Therefore, cloud vendors will continue to make high value services that are difficult to switch from, as I argued in a recent blog on multi-cloud interoperability.

It appears the MAGs are going to dominate this market for a fairly long time. But, given the rapid pace of technology disruption, nothing is certain. Moreover, having alternatives on the horizon will keep MAGs on their toes and make enterprise decisions with MAGs more balanced. Enterprises should keep investing in in-house business and technology architecture talent to ensure they can correctly architect what’s needed in the future and migrate workloads off a cloud platform when and if the time comes. Enterprises should also realize that preferring multi-cloud and actually building internal capabilities for multi-cloud are two very different things. In the long run, most enterprises will have one strategic cloud vendor and two to three others for purpose-built workloads. However, they shouldn’t be suspicious of the cloud vendors and shouldn’t resist leveraging the brilliant native services MAGs and AIO have built.

What has your experience been working with MAGs and AIO? Please share with me [email protected].

IBM Splits Into Two Companies | Blog

IBM announced this week that it is spinning off its legacy Managed Infrastructure business into a new public company, thus creating two independent companies. I highly endorse this move and, in fact, advocated it for years. IBM is a big, successful, proud organization. But it has been apparent for years that it faced significant challenges in trying to manage two very different businesses and operate within two very different operating models.

Read more in my blog on Forbes

Cloud Wars, Chapter 4: Own the “Low-Code,” Own the Client | Blog

In my series of cloud war blog posts, I’ve covered why the war among the top three cloud vendors is so ugly, the fight for industry cloud, and edge-to-cloud. Chapter 4 is about low-code application platforms. And none of the top cloud vendors – Amazon Web Services (AWS), Microsoft Azure (Azure), and Google Cloud Platform (GCP) – want to be left behind.

What is a low-code platform?

Simply put, a platform provides the run time environment for applications. The platform takes care of applications’ lifecycle as well as other aspects around security, monitoring, reliability, etc. A low-code platform, as the name suggests, makes all of this simple so that applications can be built rapidly. It generally relies on a drag and drop interface to build applications, and the background is hidden. This setup makes it easy to automate workflows, create user-facing applications, and build the middle layer. Indeed, the key reason low-code platforms have become so popular is that they enable non-coders to build applications – almost like mirroring the good old What You See is What You Get (WYSIWYG) platforms of the HTML age.

What makes cloud vendors so interested in this?

Cloud vendors realize their infrastructure-as-a-service has become so commoditized that clients can easily switch away, as I discussed in my blog, Multi-cloud Interoperability: Embrace Cloud Native, Beware of Native Cloud. Therefore, these vendors need to have other service offerings to make their propositions more compelling for solving business problems. It also means creating offerings that will drive better stickiness to their cloud platform. And as we all know, nothing drives stickiness better than an application built over a platform. This understanding implies that cloud vendors have to move from infrastructure offerings to platform services. However, building applications on a traditional platform is not an easy task. With talent in short supply, necessary financial investments, and rapid changes in business demand, enterprises struggle to build applications on time and within budget.

Enter low-code platforms. This is where cloud vendors, which already host a significant amount of data for their clients, become interested. A low-code platform that runs on their cloud not only enables clients to build applications more quickly, but also helps create stickiness because low-code platforms are notorious for “non interoperability” – it’s very difficult to migrate from one to another. But this isn’t just about stickiness. It’s one more initiative in the journey toward owning enterprises’ technology spend by building a cohesive suite of services. In GCP’s case, it’s realizing that its API-centric assets offer a goldmine by stitching together applications. For example, Apigee helps in exposing APIs from different sources and platforms. Then GCP’s AppSheet, which it acquired last year, can use the data exposed by the APIs to build business applications. Microsoft, on the other hand, is focusing on its Power platform to build its low-code client base. Combine that with GitHub and GitLab, which have become the default code stores for modern programming, and there’s no end to the innovation that developers or business leaders can create. AWS is still playing catch-up, but its launch of Honeycode has business written all over it.

What should other vendors and enterprises do?

Much like the other cloud wars around grabbing workload migration, deploying and building applications on specific cloud platforms, and the fight for industry cloud, the low-code battle will intensify. Leading vendors such as Mendix and OutSystems will need to think creatively about their value propositions around partnerships with mega vendors, API management, automation, and integration. Most vendors support deployment on cloud hyperscalers and now – with a competing offering – they need to tread carefully. Larger vendors like Pega (Infinity Platform,) Salesforce (Lightning Platform,) and ServiceNow (Now Platform) will need to support the development capabilities of different user personas, add more muscle to their application marketplaces, generate more citizen developer support, and create better integration. The start-up activity in low-code is also at a fever pitch, and it will be interesting to see how it shapes up with the mega cloud vendors’ increasing appetite in this area. We covered this in an earlier research initiatives, Rapid Application Development Platform Trailblazers: Top 14 Start-ups in Low-code Platforms – Taking the Code Out of Coding.

Enterprises are still figuring out their cloud transformation journeys. This additional complexity further exacerbates their problems. Many enterprises make their infrastructure teams lead the cloud journey. But these teams don’t necessarily understand platform services – much less low-code platforms – very well. So, enterprises will need to upskill their architects, developers, DevOps, and SRE teams to ensure they understand the impact to their roles and responsibilities.

Moreover, as low-code platforms give more freedom to citizen developers within an enterprise, tools and shadow IT groups can proliferate very quickly. Therefore, enterprises will have to balance the freedom of quick development with defined processes and oversight. Businesses should be encouraged to build simpler applications, but complex business applications should be channeled to established build processes.

Low-code platforms can provide meaningful value if applied right. They can also wreak havoc in an enterprise environment. Enterprises are already struggling with the amount of their cloud spend and how to make sense of the spend. Cloud vendors introducing low-code platform offerings into their cloud service mix is going to make their task even more difficult.

What has your experience been with low-code platforms? Please share with me at [email protected].