Cloud Field Day 13 (CFD13), part of the Tech Field Day series of events, starts on 16 February 2022. This post reviews the presenting companies and discusses what might form part of the technical presentations.
Fortinet is a company new to me, mainly as at Architecting IT we don’t cover networking. Let’s face it, most networking challenges can be fixed by stringing together a few NETGEAR hubs, right? On a serious note, it appears that Fortinet is a well-established public company with a wide range of technologies addressing security challenges across networking in the public and private cloud spaces.
As the image below shows, the enterprise data centre is no longer a physical space but a virtual environment made up of on-premises, cloud, edge/remote and even personal infrastructure. In the 1980s, access to the network was limited indirectly with proprietary protocols and physical cabling. Today, the Internet provides (and demands) widespread access across any piece of physical and virtual infrastructure.
Looking back at previous presentations, Fortinet has covered topics in mobility, security, and cloud. I’m hoping for presentations and discussions on zero trust, hybrid networking, intrusion detection and front-end security.
Kasten is a vendor of data protection solutions for Kubernetes. The company was founded in 2017 and acquired by Veeam Software in October 2020 for $150 million. Following the acquisition, the company is now branded as “Kasten by Veeam”.
Kasten K10 creates a data protection environment within a Kubernetes cluster that protects applications and data within that cluster. The data itself is typically written to an object store, either in the public cloud or on-premises. If the cluster becomes corrupted or fails, K10 offers a restore process to bring data and application definitions back to an operational state. K10 also provide disaster recovery capabilities and the option to replicate to a separate Kubernetes cluster, typically at an alternative location.
Data protection of Kubernetes applications using a solution installed within the cluster itself is an interesting choice of strategy. We believe this is the wrong long-term direction for the technology. Backup and restore needs to exist outside a Kubernetes cluster, as we discussed in this recent blog post. That being said, K10 offers some innovative ideas, such as Kanister for application-specific backup. The early acquisition by Veeam makes sense as a strategic move if the capabilities of K10 can be integrated into the broader Veeam portfolio, which would align with our thinking. You can read more on the future of Veeam in our recent post (registration required).
For CFD13, I expect we will see more on the most pressing data protection challenges of the day, including ransomware and exploiting public cloud immutable storage.
Metallic is a Commvault branded service that delivers SaaS-based data protection. Customers consume the service through an online portal, with support across Kubernetes, Office-type solutions, and cloud. You can find our last post on the development of Metallic here.
Data protection is changing as IT evolves to embrace the cloud era. Some of these changes have been quite rapid; for example, I would imagine that most IT organisations (up to a certain size) have adopted the Office 365 or Google models for email and office services. The evolving modern virtual data centre now encompasses on-premises, public cloud IaaS, public cloud SaaS, edge and mobile devices. Backup and restore has to adjust to this changing landscape and Metallic is the response from Commvault.
Why SaaS? If your primary computing platforms have moved into the cloud, why continue to run a backup infrastructure to support them? Adopting a SaaS model like Metallic pushes ongoing capacity planning, scaling, upgrade and other functionality to the vendor (in this case, Commvault). Data protection is an ideal candidate for consumption as a SaaS offering. It’s interesting that Commvault has chosen to develop this through a new persona. At the same time, this isn’t a surprise, and in my opinion is a good move.
I’m looking forward to seeing what the latest developments of the Metallic platform have to offer. Specifically, as customers migrate to Metallic from traditional Commvault, I’m keen to learn what migration offerings (if any) are available and how customers manage mixed configurations.
NetApp Inc has been a long-time presenter at Tech Field Day events (31 at the last count). The company started in 1992 as an on-premises vendor of network-attached storage, expanding into other forms of primary storage, object storage, data management, and most recently, the hybrid cloud.
A quick look back into the blog and podcast archives provides us with some insight into the recent strategy NetApp has adopted. We first discussed the data fabric with then CTO Mark Bregman in this Storage Unpacked podcast from 2016. This preview post from Cloud Field Day 3 talks about the Data Fabric, start-up acquisitions and the emergence of the NetApp SaaS portal. In the almost four years since that post was written, NetApp has worked with public cloud vendors to deploy ONTAP into AWS, Azure and GCP. The differentiator for these solutions over the competition is the native integration with existing cloud-based services.
Our thinking on NetApp is summed up in this post (registration required). Hardware needs to take a back seat as data management becomes the dominant service offering. For NetApp and its customers, this process is a multi-year transition that moves the focus from hardware “speeds and feeds” to one centred on service-based metrics. This change is well underway, positioning the company for the next 30 years of evolution.
In terms of what we can expect from CFD13, we’re likely to see some further discussion on Amazon FSx for ONTAP (see this post for details). Having discussed cloud-based file systems in this recent Storage Unpacked Podcast, we have questions on integration, data mobility and how NetApp could introduce workflow to both cloud and on-premises data. NetApp also has a push for Kubernetes support with Astra, which may also be covered (although we have no direct insight on this).
Pure Storage is a vendor of all-flash storage solutions for on-premises and cloud deployments. The company was founded as “Os76” in 2009, going public through IPO in October 2015. Initially, Pure focused on two product lines – FlashArray, a block-based storage solution for the data centre, and FlashBlade, a scale-out object and file storage platform.
In September 2020, Pure Storage acquired Portworx, a Kubernetes storage start-up, for $370 million. The Portworx Enterprise and px-backup solutions now sit in the Cloud Native Business Unit (CNBU) within the company. As the upcoming event we’re covering is a Cloud Field Day, I expect we will see presentations from the Pure cloud business unit, either focusing on Portworx or Cloud Block Store. Back in September 2021, Pure announced Fusion and Portworx Data Services. The former provides cloud-based SaaS management of FlashBlade/FlashArray, while the latter enables customers to provision typical database applications onto Portworx Enterprise deployments.
We’re hoping that the CFD13 discussion will focus on the data services offerings and how they fit into a hybrid cloud model, deploying either on-premises or onto cloud-based Kubernetes clusters. You can read more on our thoughts about the direction of Pure Storage in this post (registration required).
We have more content on Pure storage at our dedicated Pure Storage Microsite. Pure Storage is a tracked vendor for Data Storage, Data Protection, and Cloud Storage. Pure Storage is a client of Brookend Ltd.
Infrastructure as Code (IaC) – what a great concept and one that, in some form or other, we’ve been doing for the last 70 years. Yes, folks, mainframes were infrastructure as code, dividing up physical resources using an IOCP/IOCDS. In those days, though, the process certainly wasn’t automated. Over the years, we’ve developed tools and techniques for self-provisioning virtual machines and instances. I’m a big fan of automation and the concept of infrastructure as code. In recent years I’ve discussed lab automation with DHCP servers, running ESXi on Raspberry Pi, and automating tasks with APIs (here and here).
So, I’m interested to learn what Digital Rebar from RackN can do. Modern IaC, in my opinion, should provide me with the ability to deploy hardware infrastructure, virtual instances, containers, storage, networking and applications. At the same time, I want abstraction, so a single piece of code should deploy me a virtual instance in any public or private cloud. Yes, this perhaps seems like a pipe dream, but a genuinely flexible hybrid cloud won’t work unless the provisioning process can pick up a generic definition and apply it to any cloud. I’m looking forward to this presentation!
No Tech Field Day event would be complete without some discussion of Kubernetes. This looks to be where StormForge comes in, optimising Kubernetes cluster deployments. Why does this need exist?
Each decade of IT has resulted in a gradual transformation that has reduced the friction involved in deploying applications. The ease of provisioning has increased through the use of VMs, which remove the hassle of buying and deploying physical servers, containers, which remove the hassle of deploying physical VMs, and serverless, which removes the hassle of deploying anything.
In reality, applications and code still need to run somewhere. Underneath the layers of abstraction, we still have physical CPUs, memory, storage, and networking. Modern infrastructure makes the deployment process easier but also removes any requirement to be efficient. Most IT professionals will be familiar with VM sprawl, which will inevitably become an issue with containers and Kubernetes. StormForge optimises applications deployed onto Kubernetes using a mix of analysis and experimentation.
In a fully hybrid world, an application could be deployed across many pieces of infrastructure, physical and virtual. So, there’s an aspect to optimisation that has to align efficiency with cost. I’m interested to see what StormForge has to offer and how the performance, cost and efficiency metrics align.
VMware Inc needs no introduction and should be known to everyone in the technology industry. At the end of the last millennium, the company developed server virtualisation for the x86 architecture, based on solutions that initially worked on the desktop. The inefficiency of the Windows operating system, coupled with the increasing adoption of Linux, provided the perfect environment to deliver the consolidation efficiencies offered by ESXi and vSphere. Today, very few IT organisations would consider an on-premises strategy that didn’t include VMware technology.
The emergence and incredible popularity of the public cloud signalled significant risk for VMware. As we wrote in 2016, the storm clouds were brewing, and change was needed. VMware met the challenge by working with the public cloud rather than against it. The VMware vSphere platform is now available from AWS, Azure, GCP, Oracle and IBM (and smaller providers). It’s also possible to build out a VMware environment in Equinix with Metal.
Outside of the core virtualisation platform, VMware has embraced containers with Tanzu (which we reviewed as Project Pacific) and now wants to bring together the hybrid cloud message with a strategy called Cross-Cloud Services. The idea of “one management tool to rule them all” has been mooted for decades, so whether VMware can make Project Ensemble deliver the goods is, of course, debateable. At the same time, the company has to consider the long-term risk of increasing cloud adoption, so multi-cloud management has obvious attractions.
At CFD13, I’m hoping the VMware team can expand on the hybrid cloud strategy, with consideration of the continuing abstraction away from ESXi as the virtual instance hypervisor. From an analyst perspective, we’d also like to understand what difference VMware’s independence will make to the ability to support a cross-cloud model. This should be a fascinating session.
You can find more of our coverage of VMware through the dedicated Architecting IT VMware Microsite.
Copyright (c) 2007-2022 – Post #e32a – Brookend Ltd, first published on https://www.architecting.it/blog, do not reproduce without permission.