devopsy.kinsta.cloud

Tag: networking

  • What It Will Take To Bring Cloud Gaming to Fruition?

    What It Will Take To Bring Cloud Gaming to Fruition?

    The gaming industry is booming. There are 211 million people gaming in the U.S., and it’s one of the fastest-growing forms of entertainment on the planet. In the age of Twitch streaming, gaming has even upstaged traditional sports viewership. On average, young gamers between the ages of 18 and 25 spend a full hour more watching online gaming than watching traditional sports.

    Gaming has also become a way of connecting with others—65% of players say they play with others online or in person. Some software development teams even encourage gaming in the workplace to increase productivity. For this reason, there has been a massive shift toward online cooperative games, RPGs and sandboxes like Minecraft. Also, games are becoming more cloud-based, with purchases, downloads and game saves occurring in the cloud. Still, many people lack the hardware necessary to run complex games locally. All these factors are harbingers of a coming shift away from PC and console-based processing.

    I recently met with Noam Salinger, product manager, Granulate, to learn about the future of gaming and how software teams in the gaming industry can prepare. To Salinger, “cloud gaming is the next step.” The move away from consoles and toward cloud-based, remote processing will have huge ripple effects. Lowering the barrier to gaming could introduce millions of more players into the ecosystem and further balloon the industry. However, cloud gaming presents significant hurdles. The transition will require a substantial reduction in latency, 5G networking and highly efficient computing optimizations.

    Roadblocks to Cloud Gaming

    “Cloud is an integral part of any gaming company,” says Salinger. More and more functionality associated with gaming is being pushed online. There are already online game marketplaces, cloud-based account management and remote game saves. This is especially prevalent with mobile gaming. But, what about the actual processing for computationally-intensive games—could these workloads also move to the cloud soon?

    This kind of cloud-based gaming could soon come to fruition. Yet, Salinger foresees some major roadblocks to cloud gaming:

    1. The Software Components

    A complete jump to cloud gaming will require having the proper software infrastructure in place and getting vendors on board to support it. This is perhaps the biggest obstacle to cloud gaming in the software industry, said Salinger, “It’s still not mature enough.”

    2. The Business Components

    Furthermore, gaming platform companies might refrain from cloud gaming due to fears of it cannibalizing their existing hardware sales channels. On the other hand, cloud gaming presents a big opportunity to take on more of a subscription-based model, such as allowing a user to play any game they want from an extensive suite of games. Akin to streaming services like Netflix, Hulu or Peacock, we could see similar platform styles emerging in gaming. “They will have no choice but to participate at some point,” said Salinger.

    3. The Network Components

    As most gamers can attest, online games can easily suffer from network issues that negatively affect gameplay. Salinger predicts that the ubiquity of 5G networks will be necessary to propel the cloud gaming market forward. 5G could be a universal method to enable real-time connections for a broader base of end-user gamers.

    Even if software developers optimize game processing speeds, a poor network could add enough latency to data throughput to cancel out those gains. 5G could expedite this process, but it’s taking a lot longer than expected. Some estimates place 5G finally arriving at full throttle sometime in 2023, after more antennas are constructed and pre-existing physical infrastructure is upgraded.

    4. The Latency Components

    Another significant roadblock to cloud gaming is figuring out real-time remote computing on the game vendor side. “This is where the latency aspect of things really comes into play,” said Salinger. Since game clients will be making many remote calls, optimizing cloud CPU utilization will be vital to reducing the response time to requests.

    Cloud gaming will require parallel running processes that anticipate user action. Similar to the way modern web applications simultaneously trigger multiple backend requests for authentication, geolocation, user databases and so on, real-time cloud games will require many different processes running in parallel. Servers could allocate resources before requests come in, using predictive analysis of usage patterns to trigger components in advance of user behaviors.

    To enable intelligent resource allocation, Salinger describes how an OS-agnostic kernel module could be in charge of resource allocation, continuously adapting to changes in the application workload. Low-level resource allocation at the CPU level, the allocation memory, I/O, and the sequence of execution must be optimal for each specific application, he described.

    This sort of process is already being applied in production by some ad tech companies to enable real-time ad bidding. For example, by optimizing cloud CPU utilization, Perion is able to perform faster, process more responses and use fewer machines to perform the same task—thereby reducing the roundtrip time for each request. Similar principles could be used to advance cloud gaming.

    Rendering sophisticated virtual environments is highly resource-intensive, and high-performance graphics cards are expensive. Video games are set to only increase in complexity with new AR/VR modes. Cloud gaming could help respond to rising computational needs and replace the need for local computing hardware. In the process, this move could bring in more gamers, disrupt the console wars, and nullify age-old PC-versus-Mac compatibility issues.

    Moving to a cloud-based approach is also arguably more aligned with today’s fast-paced software release cadences. Pushing more frequent patches can help resolve bugs faster. But it also could aid in retaining engagement by introducing new levels and objectives continually.

    The rise of the gaming ecosystem can’t be ignored. But as it stands now, there are still significant challenges to wider penetration of cloud gaming. Cloud computing optimizations to reduce latency, mainstream adoption of 5G and significant re-platforming of software infrastructure are just some of the many issues that will need to be tackled before truly cloud-based gaming experiences advance to the next level.

  • Private and Virtual Private LTE/5G Challenges Network Visibility

    Private and Virtual Private LTE/5G Challenges Network Visibility

    Digital transformation continues to disrupt the entire IT stack by shifting nearly every paradigm, including the new blurring lines between network infrastructure. New mobile network technologies are at the cusp of adaption in the enterprise networks with a growing number of innovations poised to enable expanded and more flexible connectivity with a gradual displacement of traditional Wi-Fi networks.

    The evolution began as new unlicensed radio frequencies become a part of the 3GPP 5G Release 16 mode of operation. This new mobile mode of operation has the ability to offer the necessary technology for mobile network operators (MNOs) to operate the unlicensed spectrum into 5G networks and make some of it available for corporate networking. New standardization will enable the enterprise networks to gradually adopt these new mobile networks to their connectivity framework, in order to leverage the benefits of new modern applications and alignment to their connectivity, scale and quality of service requirements.

    With the benefits of a new mode of connectivity – in a converged fixed or pure mobile infrastructure – there are always measurable risks that require attention. Specifically, the new mode of connectivity brings potentially new security threats and operational challenges that need to be addressed.

    New Operating Modes

    It is likely that incorporating this mobile technology will require starting from new operational modes; these will be imperative to grasp the direction and the adaption of private mobile networks in reality. Industry hype is certainly driving the trend towards the mode in which the corporate IT teams will be capable to deploy and operate the new private mobile network. However, at this point in time, the more realistic option will be that full control of design, deployment and operation will be from the MNOs, as they have strong experience, knowledge and risk mitigation capabilities. This mode of operation is also incentivized through the business aspects of creating a new revenue stream for the MNOs. This higher-margin business can provide more stringent security operations, particularly in light of better understanding the vector of attacks. Such an approach provides carrier-grade operations with the commitment of a service level agreement and correlation to the required corporate business agility and continuity needed by the enterprise.

    Questions like the who, what and where of network traffic need to be answered to prevent security threats and maintain network performance and reliability—in essence, identifying and maintaining the identity of the mobile subscriber in a complex hybrid model of mobile connectivity. A corporate mobile network can be completely private, well isolated from other networks with no roaming into the public network. It can also feature a less stringent traffic policy using a virtual slice of the private network with a segmented perimeter in which certain traffic will roam into a public network and will be tracked with respect to subscriber-aware policy. This latter option is a greater challenge because of the inherent separation of data or user and control planes in mobile network architecture is virtual rather than physical. Additionally, new mobile 5G architecture includes virtualization and network slicing to offer valuable versatility and scale, but it also makes NetOps and SecOps monitoring and control even more challenging. In light of the complexities, we can assume that the favorable and practical mode of new corporate mobile networks will be outsourced to an MNO that will operate in tight synergies with corporate IT teams. Their work will be based on a defined playbook to drive a gradual evolution of architectural reference models and pragmatic learning curve for corporate “do it yourself” operations mode. Complexities multiply in cases related to IoT and manufacturing applications in which the mobile core is required to be closer or co-located with the edge to support low latency and high throughput requirements.

    5G features a separation of data/user and control planes that is an architectural advantage, particularly helpful for large IoT deployments, dynamic scalability and services and higher data rates with greater bandwidth efficiencies. It offers more distributed processing for greater scale and less susceptibility to a single point of failure. It also represents a shift from best-effort networks to ones that can provide better, more differentiated levels of services and prioritization. This separation requires a new approach for network visibility to understand who is doing what. This is an issue for security as well as networking performance.

    The Challenge of Virtualized Network Functions

    Another challenge is that 5G virtualizes in software the higher-level network functions that traditionally have been performed by physical hardware. This virtualized network slicing means a sharing of physical resources that expand the attack surface by offering the potential for a sophisticated attacker to traverse from one virtual instance to another knowing that it will be difficult for security teams to detect this lateral movement. In addition, 5G radio network deployments may include significant expansions of small cells connecting in a campus geographically distributed environment, machine-to-machine communications and dynamic connections to multiple cells. All of these present a new frontier for attackers and new challenges for IT operation teams. Organizations need visibility and understanding of these new infrastructure resources to ensure a clear design and implementation to prevent security breaches and disruption of business continuity.

    The potential for significant expansion of IoT devices on corporate networks using 5G offers important advances for measuring and monitoring devices or making everything connected. At the same time, these lean devices generally have far less endpoint controls—particularly robust access control—than traditional computing devices, and it is difficult to add such controls directly to each unit. The rise of more IoT machine-to-machine communication and away from a hub-and-spoke model adds additional complexity. A proliferation of such devices also intensifies the need for deeper visibility and more automated security to scale with these deployments without the bottleneck of human intervention for each. This adds risks to the internal IT learning curve, making it an unlikely option and again, favors outsourcing to MNO experienced teams to design, deploy and maintain the corporate private mobile networks.

  • AI Everywhere Requires MLOps and NetDevOps to Intersect

    AI Everywhere Requires MLOps and NetDevOps to Intersect

    As xOps workflows become more prevalent in IT, it’s inevitable to see overlap and intersection between them. In networking, NetDevOps has been a hot topic for the past three years. NetDevOps focuses on bringing automation, simulation and validation workflows to network operations. The goal is to reduce errors in production configuration deployment, reduce time to production and create a collaborative workflow across cross-functional teams and network team members.

    MLOps followed a similar trajectory, but with a focus on machine learning workflows. As networking evolves and machine learning and artificial intelligence (ML/AI) start to become more prevalent in all of IT, it’s inevitable that these different operational workflow methodologies start to intersect. NetDevOps evolved out of DevOps, applying lessons learned from software development workflows and applying them to network development. As the conversation shifts to ML/AI, the question currently facing many cutting-edge networking professionals in the data center space is, “Where do NetDevOps and MLOps intersect?”

    The Intersection of MLOps and NetDevOps

    Machine learning and artificial intelligence are most commonly applied to networking in three basic ways:

    • Autonomic networks
      • Intent-based configurations
      • Network devices that self-configure based on contextual information
    • Self-healing networks
      • React to network events to mitigate downtime
    • Telemetry analysis and correlation
      • Take streaming telemetry data and correlate it in a meaningful way

    All the above applications can be implemented in fairly simple ways, and many networks today have at least some variation of these technologies already applied. The challenge is that the majority of these current implementations don’t use true ML/AI. They tend to use decision trees or switch-case conditionals to create conclusions. These conclusions are reached by feeding in a pre-coded set of expected outputs. This requires large datasets, which already exist for fields like image recognition. For instance, it’s fairly easy to confirm the technology is seeing pictures of a banana; although, if the data set is, say, 100,000 images, it may take a very long time to do so manually. ML engineers tend to use existing datasets for this exact reason, but the problem is, few of these datasets exist today in the networking realm.

    Creating a solid, structured ML workflow requires a model that can be trained. This model can iterate multiple times and derive conclusions that do not need to be preprogrammed. In essence, we are trying to create these big datasets ourselves (using automation, of course) instead of relying on existing ones. But to develop such a model, certain foundations are required.

    NetDevOps Foundations

    The first foundation required for ML workflows in NetDevOps is a structured configuration model. This is mostly implemented using automation and standardized data structures. While models like OpenConfig and YANG exist, they are not comprehensive across all features commonly implemented in data centers, so most data models will have to be customized for the use case. Using this data model, and implementing it with automation technology, the ML workflow has an easy way to deploy configuration across an entire network’s fabric without having to configure each node by hand.

    The second foundation required is an environment that allows idempotent configuration iteration so the model can train itself. Idempotence refers to any function you can repeat several times without changing the final result. This is where network simulation is critical. A comprehensive network simulation platform allows the ML system to apply the standardized configuration model, and iterate over configuration evolution. That allows the application to make small configuration tweaks to allow for a nearly unlimited dataset.

    Finally, a good foundation in network MLOps requires telemetry. Historically, telemetry was used as a method to correlate logs and identify problems. Most telemetry solutions have fallen short, though, when it comes to root cause identification and action plan recommendation. These problems are much harder to solve.

    One approach could use a simulation platform to iterate configurations and architectures that eventually lead to matching production logs with simulation attempts. This methodology would train the AI model to identify which triggers affect which nodes. In this methodology, we have two independent tracks; the first to correlating logs to identify patterns and trends in the data, and the second iterating on a network architecture until the logs generated match those from production.

    This solution creates many interesting parallels to existing AI/ML solutions. For instance, iterating on a network would be equivalent to IBM’s DeepBlue learning chess, or AlphaGo learning to play Go. The model trains by playing games, in simulation, against themselves, over and over. With every iteration, they learn something new about how each decision influences an outcome. Given enough iterations, the machines become unbeatable. Matching logs from simulation to production is similar to AI/ML in facial and voice recognition, too. In recognition software, pattern matching is used to train a model that becomes increasingly efficient at identifying images the more data it is fed.

  • The Biggest Motivators and Barriers for SD-WAN Adoption in 2021

    The Biggest Motivators and Barriers for SD-WAN Adoption in 2021

    The SD-WAN market has exploded in 2020, driven by the need to support increasingly remote workforces, the emergence of the AI-driven WAN and the requirement to deliver branch performance that matches campus networks. As enterprises look at the year ahead, support for the distributed workforce remains a top priority, and SD-WAN is a crucial technology that businesses need to invest in to thrive in the “new normal.”

    As we head into 2021, these are the biggest motivators for the enterprise to adopt SD-WAN:

    • Increased demand for robust video quality: The days of in-person conferences and meetings have been traded with back-to-back Zoom calls. Though lags and connectivity issues are understandable, they can be frustrating and ultimately lead to decreased productivity. Experiencing these issues has increased the demand for more robust and higher bandwidth network connectivity, which is critical for video communications. This increasing demand is expected to continue to rise—although some businesses are planning to return to the office, many are adopting a hybrid work model that will require at last part-time remote work.
    • Boosting bandwidth without breaking the bank: Outside of video calls, the surge in remote work has increased pressure on overall network bandwidth. The solution for organizations is, of course, to add more bandwidth, but doing so can be expensive. SD-WAN is an attractive option for increased bandwidth pressure because it’s a solution that can meet bandwidth pressure demands while remaining cost-effective.
    • Centrally operated networks needed for distributed sites: Once enterprise video quality and bandwidth needs are fulfilled, the next step is to ensure that IT teams have the resources they need to effectively operate remote networks and deliver a great user experience from client to cloud. SD-WAN is a useful IT tool for network management because it provides full visibility into what’s happening across the network. And with the recent emergence of AI-driven WAN solutions, IT teams have more tools than ever before to get insight into user experience, enabling them to proactively and remotely resolve issues.

    Despite the many benefits of SD-WAN, there are still a number of barriers that are keeping enterprises from adopting this technology, including:

    • Concerns about quality of service (QoS) for applications across WAN: With SD-WAN solutions shifting traffic away from dedicated private circuits towards the public internet to reduce cost, businesses are often concerned about their ability to assure QoS for their critical applications. These businesses should look for solutions that provide insights into the experience from client to cloud as well as leverage data science and AI to detect anomalies. They should also automate the root cause determination and resolution through self-driving or IT actions.
    • Lack of insight into user experience: Many SD-WAN solutions lack the insight into both user and app experience from wireless to wired to WAN, whether due to the use of link-based VPN connections rather than session-based protection and/or the lack of rich streaming telemetry from the network devices to a modern cloud-based AI system. New AI-driven solutions and session-based technologies are now available that can improve this insight to simplify network operations and assure great user experiences from client to cloud.
    • Securing remote sites from potential security threats: By further extending the threat surfaces with a more distributed workforce, many organizations are concerned about the potential security threats that a distributed network could introduce. To mitigate this, it is critical to look for solutions that embrace a connected security paradigm that delivers insight into threats across the network while automating mitigation and enforcing security policies at the edge.
    • Limited technical skills and knowledge: A common perceived barrier to SD-WAN adoption is the concern that there is a significant learning curve for the technology and not enough existing skills to plan and manage it. This is a valid concern, though there are solutions. First, some SD-WAN providers offer managed services and can oversee deployment for customers through an outsourced or hybrid IT model. Second, organizations can invest in closing the skills gap via training. Though the latter option can be more time-consuming, investing in employees to learn the skills can help enterprises save money in the long run and move the industry forward by up-leveling important technical skills that are increasing in demand.

    Though there are notable barriers to SD-WAN adoption, there are also solutions for each one. If enterprises can overcome these barriers, implementing SD-WAN has the potential to improve scale, bring agility, unify policy and boost security efficacy through a collapsed and integrated AI-driven network. Deploying SD-WAN can be a significant investment for companies, but in the long run, the technology is an important step in moving enterprises into the next generation of networking.

  • Aruba Aims to Centralize Network Management

    Aruba Aims to Centralize Network Management

    Aruba, a unit of Hewlett-Packard Enterprise (HPE), unveiled orchestration software that makes it possible to centralize the management of network switches in the data center, across an extended campus and at the network edge.

    In addition, Aruba is adding a family of five 1/10/25/40/100 GbE switches to its portfolio that are designed to be deployed in the data center or network edge.

    Finally, Aruba is making available integrations for HPE ProLiant DL and DX servers as well as SimpliVity, Nimble, Synergy, Cray Shasta, Cray ClusterStor, SAP HANA, VMware and Nutanix platforms.

    William Choe, vice president of product management for Aruba, said Aruba Fabric Composer can also be coupled with HPE GreenLake, a managed service HPE provides to remotely manage HPE servers on behalf of customers.

    Based on a graphical interface, Aruba Fabric Composer is designed to make it possible for IT generalists to provision any type of switch, Choe said. The goal is to reduce the total cost of networking by enabling IT teams to centralize the management of switches regardless of whether they are a data center, campus or edge computing environment.

    With the rise of edge computing, the number of network switches that are being employed across a distributed computing environment is increasing. The challenge IT teams face is they can’t easily dispatch personnel to physically provision switches in remote locations at a time, as most IT staff members are working from home to help combat the spread of the COVID-19 pandemic.

    Aruba Fabric Composer, which is designed to be accessed via a VMware console, eliminates the need to manage enterprise networks in a piecemeal fashion, added Choe.

    Arguably, the centralization of network management is long overdue. Too many IT organizations are still managing each individual switch via a command-line interface (CLI) for each box. That not only makes managing a large network cumbersome, but it also creates a rigid IT environment that is unable to adroitly respond to rapid changes being made to the application environment. At a time when many organizations are trying to accelerate the deployment of applications needed to drive digital business transformation initiatives, legacy approaches to networking are hampering the modernization of IT.

    There are other networking equipment providers working to centralize the management of switches. However, it remains to be seen whether IT organizations are prioritizing the centralization of network management during the economic downturn brought on by the pandemic. But there is an absolute need to increase network bandwidth, so one way to justify an upgrade to the network is to make it less expensive to operate.

    As part of that effort, many organizations are also starting to converge network operations with best DevOps practices using application programming interfaces (APIs) and overlays such as service meshes to programmatically manage network underlays. Regardless of approach, the days when networks were managed by a dedicated team of specialists are on the wane.

  • Cisco CEO: Developers Need More Infrastructure Control

    Cisco CEO: Developers Need More Infrastructure Control

    Cisco Systems CEO Chuck Robbins said this week that the company’s No. 1 priority going forward is to enable IT organizations to better manage application experience via abstraction layers that make to possible for developers to programmatically manage infrastructure.

    Speaking at its annual Cisco Partner Summit, which occurred online, Robbins said Cisco is working toward making it possible for developers to not only programmatically provision IT infrastructure but also impact how services are delivered using application networking services the company is currently developing.

    Over time, the entire Cisco portfolio will either be cloud-delivered or cloud-managed as Cisco increasingly delivers its technology as a service, he said. As part of that effort, Cisco is now exposing application programming interfaces (APIs) across its portfolio.

    At the core of the initiative are investments Cisco is making in AppDynamics, ThousandEyes and its Intersight IT operations management platform. Those tools will be employed to provide insights and observability across a diverse application environment that increasingly includes cloud-native applications running on Kubernetes clusters, Robbins said.

    In effect, Cisco is signaling a more aggressive transition toward managing infrastructure as code inside and out of the cloud. This week Cisco announced it has added an Intersight Workload Optimizer to optimize infrastructure consumption along with adding support for dynamic provisioning of Kubernetes clusters via Cisco Intersight. The company also added a Cisco Nexus dashboard to provide visibility into networks spanning multiple clouds, on-premises IT environments and edge computing platforms, and expanded the Cisco Identity Services Engine to now secure a wider variety of internet of things (IoT) endpoints.

    Cisco, like every other provider of IT infrastructure, is shifting toward a cloud operating model that promises to enable IT teams to manage on-premises IT infrastructure in the same way IT teams consume public cloud resources. As part of that effort, Robbins noted Cisco has been disaggregating all the software functionality it provides from hardware whenever possible, which will provide IT teams with more flexibility and control over highly distributed IT environments.

    The COVID-19 pandemic has only accelerated a transition to managing IT as service, as organizations rely more on applications to engage their end customers, he added.

    Of course, Cisco is not the only provider of IT infrastructure with similar ambitions. Not only is networking software is becoming accessible via APIs, but it’s also now possible to deploy a software-defined network on commodity processors almost anywhere. That transition is also the primary reason that network operations are increasingly being folded into DevOps teams seeking to manage compute, storage and networking as code across an extended enterprise.

    It may be a while yet before network and storage operations are fully assimilated into DevOps processes. However, it’s now no longer a question of if as much as how long before most IT organizations achieve that goal.

  • Cisco Revamps Portfolio at the Network Edge

    Cisco Revamps Portfolio at the Network Edge

    Cisco today announced a series of Catalyst offerings that combine the functionality of a secure software-defined wide area network (SD-WAN) and a router in a single platform.

    Archana Khetan, senior director of product management for enterprise routing and SD-WAN infrastructure within the Intent Based Networking Group at Cisco, said the Catalyst 8000 series is made up of both physical devices and virtual appliances that eliminate the need for organizations to deploy separate SD-WAN and routers at the network edge.

    The high end of the Catalyst 8500 platform line is based on a third-generation Quantum Flow Processor, a proprietary ASIC that Cisco employs in its networking equipment. It is designed for data center, colocation and aggregation sites and, for the first time for Cisco, integrates 40G and 100G Ethernet ports.

    The Catalyst 8300 Series Edge Platforms are designed for branch offices that are transitioning from relying solely on routers to embracing SD-WANs. In addition to providing 4X better performance than legacy Cisco routers, the Catalysts 8300 series is designed to allow organizations to insert modules that are compatible with existing Cisco’s ISR 4400 routers. Compute and switching capabilities can also be added via a UCS-E Series blades and UADP-based switch modules can be added to create a software-defined branch office capable of running microservices-based applications based on containers on x86 processors. Cisco also unveiled today a Catalyst Cellular Gateway, for organizations that want to connect branch offices over 4G LTE wireless networks with support for 5G networks planned.

    Finally, Cisco has added Catalyst 8000V Edge Software, which enables IT teams to deploy Cisco networking software on a public cloud. While most public clouds already provide networking services, Catalyst 8000V Edge Software provides more advanced networking and security capabilities that can be managed as an extension of an existing enterprise network, said Khetan.

    Each member of the Catalyst 8000 series has been integrated with Cisco Umbrella to provide a means of managing networking and security policies via the cloud, while Cisco SD-WAN Cloud OnRamp integrates with cloud services from Amazon Web Services (AWS), Google and Microsoft to provide visibility into those environments.

    In addition, each platform provides access to a programmable application programming interface (API) library as part of Cisco’s ongoing effort to meld network operations and DevOps workflows.

    While Cisco remains the dominant provider of networking equipment in the enterprise, the rise of SD-WAN platforms that essentially replace routers in branch offices has created a slew of competitors that are gaining footholds in branch offices at the expense of Cisco. With the launch of the Catalyst 8000 series, Cisco is aiming to bridge the gap between routers and SD-WAN appliances within a single platform that promises to be easier to manage via the cloud.

    It’s not clear to what degree that approach will resonate with enterprise IT environments, but it’s clear Cisco is now more than ready to strike back.

  • Kong Previews Application Connectivity Service

    Kong Previews Application Connectivity Service

    Kong Inc., at its online Kong Summit 2020 conference, unfurled Kong Konnect, a cloud service through which developers can leverage application programming interfaces (APIs) and microservices to integrate distributed applications.

    Available as a private beta, Kong Konnect promises to make it easier for developers to create complex workflows spanning API gateways, Kubernetes Ingress controllers and service mesh runtimes.

    Company CTO Marco Palladino said Kong Connect makes it possible for developers to consume integration runtimes as a service via an interface that allows them to invoke only the specific services they require.

    Scheduled to be generally available early next year, Kong Konnect is based on a ServiceHub platform that creates a single source of truth for all services via a catalog that any organization has developed. Kong Konnect also provides access to a runtime manager that enables self-service provisioning based on roles in addition to enabling IT teams to manage Kong runtimes across hybrid cloud computing environments.

    Kong Konnect also integrates the company’s API Gateway, Ingress Controller and Kong Mesh, a service mesh based on open source Envoy proxy software, and Kong Vitals, a tool for monitoring Kong services and transactions in real-time. Kong Konnect is also integrated with Insomnia to design APIs using a tool that integrates with any Git-based repository and the Kong Developer Portal, through which developers can publish their own services.

    Palladino said the goal is to provide a turnkey service that automatically provides all the Layer 4 through 7 networking services that developers require to integrate both monolithic applications running on virtual machines as well as emerging microservices-based applications running on Kubernetes clusters. Kong Konnect makes available universal service connectivity on-demand via a single API to any class of developer, he said.

    As application environments become more distributed, developers are spending more time than ever on basic connectivity issues. The company is making a case for a cloud service that unifies all those services under a common control plane it manages on behalf of developers. That service will then free up developers to spend more of their time writing code versus integrating application services, said Palladino.

    That capability will be especially critical as IT organizations begin to deploy more applications on edge computing platforms that need to be integrated with a wide array of application services running on public clouds and on-premises IT environments, added Palladino.

    It’s too early to say to what degree IT organizations may be ready to rely on a service versus continuing to manage connectivity themselves. In effect, Kong is making a case for abstracting all the underlying complexity of connectivity and networking using an API to invoke everything from a simple API gateway to a service mesh designed to integrate hundreds of microservices.

    Regardless of the path chosen, the one thing that is clear is a new application connectivity as a service era has finally arrived.

  • Making the Move to Edge Computing? Consider This 

    Making the Move to Edge Computing? Consider This 

    Edge computing has exciting potential, but there are certain considerations to thrive in today’s competitive landscape

    Data is king, especially in the current era of the data-driven business, where a data mindset and analytical capabilities are significant factors in determining future success. With so much data to be mined and monetized, more companies are recognizing the competitive advantages of edge computing, which can also jump-start the journey to becoming an autonomous digital enterprise. That said, a variety of security and infrastructure challenges may prevent some IT professionals from making the move, according to a recent survey by Volterra and Propeller Insights.

    Distributed models such as edge computing have always been attractive because of their ability to minimize network latency issues with large data and put essential functions closer to the technology consumer. Now, they’re even more attractive with new compute and application architectures and so many new data sources—network-connected and internet of things (IoT) devices, social media, customer engagement systems, transactional systems and systems of record—dramatically expanding the volume of data and changing the nature of the data collected. Edge computing also helps solve the traditional issues of optimizing for network capacity and performance and addresses new areas such as the localized pre-processing of all data sources (telemetry, video, mobile, etc.).

    It’s no surprise, then, that 54% of the IT leaders surveyed said they prefer edge to the cloud for operators who need to control and analyze data for local use cases, and to avoid the latency that can occur when sending edge data to public cloud-based apps. By moving certain workloads to the edge of the network instead, devices spend less time on the cloud, react faster to local changes, and operate reliably–even in extended offline periods.

    Edge computing still has its challenges and difficulties to overcome. Among them are key concerns such as a lack of resources or time for long-term upkeep of application and infrastructure, and managing distributed clusters as siloed instances rather than a single resource.

    Taking IT to the Edge – Edge Computing, That Is

    Despite the roadblocks, the value of edge computing is clear. This is why many companies are including edge computing models in their business plan. However, the performance, control and governance challenges that come with distributed models of compute, applications and data should be considered carefully beforehand.

    Innovative, data-driven companies that want to stay on the cutting edge and are planning an edge migration should develop a plan and long-term vision for their mix of data center, cloud and edge resources that consider their current and future use cases. Partnerships will also prove useful. By leveraging multi-vendor, commercial off-the-shelf solutions, enterprises can be more successful yielding value from edge computing than they would with single-vendor niche products. An ecosystem of seamlessly integrated solutions will help companies harness data and translate information into insight and actions that can optimize assets.

    As today’s enterprises seek to become more data-driven and evolve into the autonomous digital enterprises of tomorrow, it’s important to think about edge use cases as part of an overall automation strategy. Companies should consider how edge applications and distributed architectures will support them from a capacity, monitoring, deployment and service management perspective.

    Finally, edge computing’s advantages also come with the risk of security vulnerabilities if they’re not addressed from the outset, since the “attack surface” of a company’s IT estate becomes much larger. Organizations must deploy a multi-layer security approach that aligns with their edge initiatives while factoring in service-level needs to ensure assets are adequately protected.

    Edge computing has exciting potential, and the enterprises that make strategic investments and take a holistic, outlook-driven, secure approach to migration will thrive in tomorrow’s competitive landscape and years into the future.

  • Juniper Networks Brings AI to SD-WANs

    Juniper Networks Brings AI to SD-WANs

    Juniper Networks today announced it can now apply the same artificial intelligence (AI) engine it uses to automate the management of wireless local area networks (LANs) to software-defined wide area networks (SD-WANs).

    Based on an AI engine that Juniper Networks gained when it acquired Mist Systems, the Juniper Mist WAN Assurance cloud service streams telemetry data gathered from Juniper SRX devices to an AI engine that advises IT teams on how to best set WAN service levels as part of a larger transition to intent-based networking.

    As part of that initiative, Juniper Networks is also providing access to a Marvis conversational interface originally developed by Mist to enable IT teams to verbally issue commands, replacing the traditional command-line interface (CLI).

    Christian Gilby, director of product marketing for Juniper Networks, said as the number of remote branch offices that IT organizations need to manage increases at a time when travel is limited, the need to apply AI to SD-WANs has never been more apparent.

    Rather than return to central offices in the wake of the COVID-19 pandemic, Gilby noted, many organizations are shifting staff to smaller remote offices to enable collaboration while continuing to minimize potential exposure to anyone who might be infected.

    The challenge is that many IT teams tasked with managing the WANs that connect those remote offices aren’t growing during the downturn caused by the pandemic. AI provides a way to troubleshoot and optimize WAN performance without having to increase network operations headcount, Gilby said.

    Platforms such as the Juniper Mist WAN Assurance cloud service provides the added benefit of never forgetting anything they learn or taking a day off.

    Gilby said it’s not likely AI systems will replace the need for network operations teams anytime soon. However, Juniper Mist WAN Assurance is part of a larger wave of AIOps technologies that are augmenting existing IT staff. It may take a while for them to trust the optimizations being made by AIOps platforms, but the more data those systems collect, the more accurate they become over time.

    In the case of a cloud service, the machine learning algorithms that drive the AI models are being trained continually using anonymized data collected from thousands of systems.

    The launch of Juniper Mist WAN Assurance comes at a time when more organizations are looking to embrace SD-WANs as a way to more easily scale the delivery of network services when employees are working from home for the foreseeable future. In many cases, IT teams are now dropping Mist wireless routers into home offices to extend the reach of SD-WAN services in a way that can be managed and optimized using the same core AI technologies, said Gilby. In its most recent quarter, Juniper Networks reported a 170% rise in orders year-over-year for Mist wireless LANs.

    It remains to be seen how much or how quickly networking operations will be transformed by AI. However, at this point, that transformation is all but inevitable.