Hyper-Converged Infrastructure, Windows Systems and Network
Article | July 11, 2023
Pacific Electric Wire & Cable Co. (PEWC) is a manufacturer in Taiwan with subsidiaries in China, Singapore, Thailand, and Australia. Like many companies, they had been facing the looming change over to SAP HANA. They were ready to switch over from their older SAP software and take advantage of SAP HANA apps and databases. They also had a goal of speeding up operational analytics and insights. But with the change to HANA, they needed all new infrastructure, certified by SAP, to support it.
Read More
Application Infrastructure, Application Storage
Article | July 19, 2023
Adopting DevOps and CD in IaaS environments is a strategic imperative for organizations seeking to achieve agility, competitiveness, and customer satisfaction in their software delivery processes.
Contents
1. Introduction
2. What is IaaS Virtualization?
3. Virtualization Techniques for DevOps and Continuous Delivery
4. Integration of IaaS with CI/CD Pipelines
5. Considerations in IaaS Virtualized Environments
5.1. CPU Swap Wait
5.2. CPU System/Wait Time for VKernel:
5.3. Memory Balloon
5.4.Memory Swap Rate:
5.5. Memory Usage:
5.6. Disk/Network Latency:
6. Industry tips for IaaS Virtualization Implementation
6.1. Infrastructure Testing
6.2. ApplicationTesting
6.3. Security Monitoring
6.4. Performance Monitoring
6.5. Cost Optimization
7. Conclusion
1. Introduction
Infrastructure as a Service (IaaS) virtualization presents significant advantages for organizations seeking to enhance their agility, flexibility, and speed to market within the DevOps and continuous delivery frameworks. Addressing the associated risks and challenges is crucial and can be achieved by employing the appropriate monitoring and testing techniques, enlisted further, in this blog.
IaaS virtualization allows organizations to provision and de-provision resources as needed, eliminating the need for long-term investments in hardware and data centers. Furthermore, IaaS virtualization offers the ability to operate with multiple operating systems, databases, and programming languages, empowering teams to select the tools and technologies that best suit their requirements.
However, organizations must implement comprehensive testing and monitoring strategies, ensure proper security and compliance controls, and adopt the best resource optimization and management practices to leverage the full potential of virtualized IaaS. To achieve high availability and fault tolerance along with advanced networking, enabling complex application architectures in IaaS virtualization, the blog mentions five industry tips.
2. What is IaaS Virtualization?
IaaS virtualization involves simultaneously running multiple operating systems with different configurations. To run virtual machines on a system, a software layer known as the virtual machine monitor (VMM) or hypervisor is required.
Virtualization in IaaS handles website hosting, application development and testing, disaster recovery, and data storage and backup. Startups and small businesses with limited IT resources and budgets can benefit greatly from virtualized IaaS, enabling them to provide the necessary infrastructure resources quickly and without significant capital expenditures.
Virtualized IaaS is a potent tool for businesses and organizations of all sizes, enabling greater infrastructure resource flexibility, scalability, and efficiency.
3. Virtualization Techniques for DevOps and Continuous Delivery
Virtualization is a vital part of the DevOps software stack. Virtualization in DevOps process allows teams to create, test, and implement code in simulated environments without wasting valuable computing resources. DevOps teams can use the virtual services for thorough testing, preventing bottlenecks that could slow down release time. It heavily relies on virtualization for building intricate cloud, API, and SOA systems. In addition, virtual machines benefit test-driven development (TDD) teams that prefer to begin their troubleshooting at the API level.
4. Integration of IaaS with CI/CD Pipelines
Continuous integration is a coding practice that frequently implements small code changes and checks them into a version control repository. This process not only packages software and database components but also automatically executes unit tests and other tests to provide developers with vital feedback on any potential breakages caused by code changes.
Continuous testing integrates automated tests into the CI/CD pipeline. For example, unit and functionality tests identify issues during continuous integration, while performance and security tests are executed after a build is delivered in continuous delivery. Continuous delivery is the process of automating the deployment of applications to one or more delivery environments.
IaaS provides access to computing resources through a virtual server instance, which replicates the capabilities of an on-premise data center. It also offers various services, including server space, security, load balancing, and additional bandwidth. In modern software development and deployment, it's common to integrate IaaS with CI/CD pipelines. This helps automate the creation and management of infrastructure using infrastructure-as-code (IAC) tools. Templates can be created to provision resources on the IaaS platform, ensuring consistency and meeting software requirements. Additionally, containerization technologies like Docker and Kubernetes can deploy applications on IaaS platforms.
5. Considerations in IaaS Virtualized Environments
5.1. CPU Swap Wait
The CPU swap wait is when the virtual system waits while the hypervisor swaps parts of the VM memory back in from the disk. This happens when the hypervisor needs to swap, which can be due to a lack of balloon drivers or a memory shortage. This can affect the application's response time. One can install the balloon driver and/or reduce the number of VMs on the physical machine to resolve this issue.
5.2. CPU System/Wait Time for VKernel
Virtualization systems often report CPU or wait time for the virtualization kernel used by each virtual machine to measure CPU resource overhead. While this metric can't be directly linked to response time, it can impact both ready and swap times if it increases significantly. If this occurs, it could indicate that the system is either misconfigured or overloaded, and reducing the number of VMs on the machine may be necessary.
5.3. Memory Balloon
Memory ballooning is a memory management technique used in virtualized IaaS environments. It works by injecting a software balloon into the VM's memory space. The balloon is designed to consume memory within the VM, causing it to request more memory from the hypervisor. As a result, if the host system is experiencing low memory, it will take memory from its virtual infrastructures, thus negatively affecting the guest's performance, causing swapping, reduced file-system buffers, and smaller system caches.
5.4. Memory Swap Rate
Memory swap rate is a performance metric used in virtualized IaaS environments to measure the amount of memory being swapped to disk. When the swap rate is high, it leads to longer CPU swap times and negatively affects application performance. In addition, when a VM is running, it may require more memory than is physically available on the server. In such cases, the hypervisor may use disk space as a temporary storage area for excess memory. Therefore, to optimize, it is important to ensure that VMs have sufficient memory resources allocated.
5.5. Memory Usage
Memory usage refers to the amount of memory being used by a VM at any given time. Memory usage is assessed by analyzing the host level, VM level, and granted memory. When memory usage exceeds the available physical memory on the server, the hypervisor may use disk space as a temporary storage area for excess memory, leading to performance issues. The disparity between used and granted memory indicates the overcommitment rate, which can be adjusted through ballooning.
5.6. Disk/Network Latency
Some virtualization providers provide integrated utilities for assessing the latency of disks and network interfaces utilized by a virtual machine. Since latency directly affects response time, increased latency at the hypervisor level will also impact the application. An excessive amount of latency indicates the system is overloaded and requires reconfiguration. These metrics enable us to monitor and detect any negative impact a virtualized system might have on our application.
6. Industry tips for IaaS Virtualization Implementation
Testing, compliance management and security arecritical aspects of managing virtualized IaaS environments . By implementing a comprehensive strategy, organizations ensure their infrastructure and applications' reliability, security, and performance.
6.1. Infrastructure Testing
This involves testing the infrastructure components of the IaaS environment, such as the virtual machines, networks, and storage, aiming to ensure the infrastructure is functioning correctly and that there are no performance bottlenecks, security vulnerabilities, or configuration issues. Testing the virtualized environment, storage testing (testing data replication and backup and recovery processes), and network testing are some of the techniques to be performed.
6.2. Application Testing
Applications running on the IaaS virtual environment should be thoroughly tested to ensure they perform as expected. This includes functional testing to ensure that the application meets its requirements and performance testing to ensure that the application can handle anticipated user loads.
6.3. Security Monitoring
Security monitoring is critical in IaaS environments, owing to the increased risks and threats. This involves monitoring the infrastructure and applications for potential security threats, vulnerabilities, or breaches. In addition, regular vulnerability assessments and penetration testing help identify and address potential security issues before they become significant problems.
6.4. Performance Monitoring
Performance monitoring is essential to ensuring that the underlying infrastructure meets performance expectations and has no performance bottlenecks. This comprises monitoring metrics such as CPU usage, memory usage, network traffic, and disk utilization. This information is used to identify performance issues and optimize resource usage.
6.5. Cost Optimization
Cost optimization is a critical aspect of a virtualized IaaS environment with optimized efficiency and resource allocation. Organizations reduce costs and optimize resource usage by identifying and monitoring usage patterns and optimizing elastic and scalable resources. It involves right-sizing resources, utilizing infrastructure automation, reserved instances, spot instances (unused compute capacity purchased at a discount), and optimizing storage usage.
7. Conclusion
IaaS virtualization has become a critical component of DevOps and continuous delivery practices. To rapidly develop, test, and deploy applications with greater agility and efficiency by providing on-demand access to scalable infrastructure resources to Devops teams, IaaS virtualization comes into picture. As DevOps teams continue to seek ways to streamline processes and improve efficiency, automation will play an increasingly important role. Automated deployment, testing, and monitoring processes will help reduce manual intervention and increase the speed and accuracy of development cycles. In addition, containers will offer a lightweight and flexible alternative to traditional virtualization, allowing DevOps teams to package applications and their dependencies into portable, self-contained units that can be easily moved between different environments. This can reduce the complexity of managing virtualized infrastructure environments and enable greater flexibility and scalability. By embracing these technologies and integrating them into their workflows, DevOps teams can achieve greater efficiency and accelerate their delivery of high-quality software products.
Read More
Application Storage, Data Storage
Article | July 12, 2023
The success of 5G technology is a function of both the infrastructure that supports it and the ecosystems that enable it. Today, the definitive focus in the 5G space is on enterprise use cases, ranging from dedicated private 5G networks to accessing edge compute infrastructure and public or private clouds from the public 5G network. As a result, vendor-neutral multitenant data center providers and their rich interconnection capabilities are pivotal in helping make 5G a reality. This is true both in terms of the physical infrastructure needed to support 5G and the ability to effectively connect enterprises to 5G.
Industry experts expect 5G to enable emerging applications such as virtual and augmented reality (AR/VR), industrial robotics/controls as part of the industrial internet of things (IIoT), interactive gaming, autonomous driving, and remote medical procedures. These applications need a modern, cloud-based infrastructure to meet requirements around latency, cost, availability and scalability. This infrastructure must be able to provide real-time, high-bandwidth, low-latency access to latency-dependent applications distributed at the edge of the network.
How Equinix thinks about network slicing
Network slicing refers to the ability to provision and connect functions within a common physical network to provide the resources necessary to deliver service functionality under specific performance constraints (such as latency, throughput, capacity and reliability) and functional constraints (such as security and applications/services). With network slicing, enterprises can use 5G networks and services for a wide variety of use cases on the same infrastructure.
Providing continuity of network slices with optimal UPF placement and intelligent interconnection
Mobile traffic originates in the mobile network, but it is not contained to the mobile network domain, because it runs between the user app on a device and the server workload on multi-access edge compute (MEC) or on the cloud. Therefore, to preserve intended characteristics, the slice must be extended all the way to where the traffic wants to go. This is why we like to say “the slicing must go on.”
The placement of network functions within the slice must be optimized relative to the intended traffic flow, so that performance can be ensured end-to-end. As a result, organizations must place or activate the user plane function (UPF) in optimal locations relative to the end-to-end user plane traffic flow.
We expect that hybrid and multicloud connectivity will remain a key requirement for enterprises using 5G access. In this case, hybrid refers to private edge computing resources (what we loosely call “MEC”) located in data centers—such as Equinix International Business Exchange™ (IBX®) data centers—and multicloud refers to accessing multiple cloud providers from 5G devices. To ensure both hybrid and multicloud connectivity, enterprises need to make the UPF part of the multidomain virtual Layer 2/Layer 3 interconnection fabric.
Because a slice must span multiple domains, automation of UPF activation, provisioning and virtual interconnection to edge compute and multicloud environments is critical.
Implementing network slicing for interconnection of core and edge technology
Equinix partnered with Kaloom to develop network slicing for interconnection of core and edge (NICE) technology within our 5G and Edge Technology Development Center (5G ETDC) in Dallas. NICE technology is built using cloud-native network fabric and high-performance 5G UPF from Kaloom. This is a production-ready software solution, running on white boxes built with P4 programmable application-specific integrated circuits (ASICs), allowing for deep network slicing and support for high-performance 5G UPF with extremely fast data transfer rates.
With NICE technology in the 5G ETDC, Equinix demonstrates:
5G UPF deployment/activation and traffic breakout at Equinix for multiple slices.
Software-defined interconnection between the 5G core and MEC resources from multiple providers.
Software-defined interconnection between the 5G core and multiple cloud service providers.
Orchestration of provisioning and automation of interconnection across the 5G core, MEC and cloud resources.
Architecture of NICE technology in the Equinix 5G ETDC
The image above shows (from left to right):
The mobile domain with radio access network (RAN), devices (simulated) and mobile backhaul connected to Equinix.
The Equinix domain with:
Equinix Metal® supporting edge computing servers and a fabric controller from Kaloom.
Network slicing fabric providing interconnection and Layer 2/Layer 3 cloud-native networking to dynamically activate UPF instances/interfaces connected with MEC environments and clouds, forming two slices (shown above in blue and red).
Equinix Fabric™ and multicloud connectivity.
This demonstrates the benefit of having the UPF as a feature of the interconnection fabric, effectively allowing UPF activation as part of the virtual fabric configuration. This ultimately enables high-performance UPF that’s suitable for use cases such as high-speed 5G fixed wireless access.
Combining UPF instances and MEC environments into an interconnection fabric makes it possible to create continuity for the slices and influence performance and functionality. Equinix Fabric adds multicloud connectivity to slices, enabling organizations to directly integrate network slicing with their mobile hybrid multicloud architectures.
Successful private 5G edge deployments deliver value in several ways. Primarily, they offer immediate access to locally provisioned elastic compute, storage and networking resources that deliver the best user and application experiences. In addition, they help businesses access a rich ecosystem of partners to unlock new technologies at the edge.
Secure, reliable connectivity and scalable resources are essential at the edge. A multivendor strategy with best-of-breed components complemented by telemetry, advanced analytics with management and orchestration—as demonstrated with NICE in Equinix data centers—is a most effective way to meet those requirements. With Equinix’s global footprint of secure, well-equipped facilities, customers can maximize benefits.”
- Suresh Krishnan, CTO, Kaloom
Equinix and its partners are building the future of 5G
NICE technology is just one example of how the Equinix 5G and Edge Technology Development Center enables the innovation and development of real-world capabilities that underpin the edge computing and interconnection infrastructure required to successfully implement 5G use cases. A key benefit of the 5G ETDC is the ability to combine cutting-edge innovations from our partners like Kaloom with proven solutions from Equinix that already serve a large ecosystem of customers actively utilizing hybrid multicloud architectures.
Read More
Article | May 4, 2020
Enterprises running hypervisors on hyper-converged infrastructure (HCI) systems typically have backup options available to them that are not available to those running on generic hardware. Such customers may also have additional backup challenges depending on the HCI vendor and hypervisor they have chosen. Let’s take a look.
Read More