Corporate IT is a key pillar for all of today’s industries, and data centers are at its very core:
Being the heart of the network, the infrastructure of any DC must guarantee flexibility and scalability as well as stability and maximum availability – all at the same time. Let us support you in meeting this challenge!
NEXT-GEN DATA CENTER
Xantaro NG Data Center Solutions – bridging the gap between digital transformation and core business
Many companies shy away from outsourcing their data or in fact their entire IT system to cloud service providers. Instead, they rely on solutions in their own data centers.
But digital transformation and its cutting-edge applications come with certain requirements. A number of things are crucial to reliably map all business processes: Bandwidth, performance, comprehensive access and low latency play just as important a role as network management and infrastructure security.
With its Next-Gen Data Center Solutions, Xantaro offers standardised, future-orientated modules and solutions aimed at modernising data center architecture.
Xantaro Next-Gen Data Center Solutions:
The latest Generation in Data Center Architectures
To help companies curb operational expenses upon setting up modern data centers and simplify the use of cloud technologies, Xantaro has standardised the design, implementation and operation of innovative data center architectures in vendor-neutral blueprints.
Thanks to continuous engineering and automated testing, new hardware and software is qualified on a regular basis. Incidents in the installed base as well as vulnerabilities announced by vendors are recorded in the problem management and incorporated both into support measures and the further development of the solution.
Flexible modules for the design, integration and operation of modern data centers
Based on our know-how and experience in the implementation and operation of a wide range of high-availability networks and from creating synergies out of various solutions implemented for several customers, we have developed blueprints for future-proof data center architectures.
The result: highly individual, vendor-neutral solutions that meet the exact requirements of scalability and flexibility as well as stability and availability.
The idea: “Divide & Conquer”
- We divide the network into layers
Transport layer (Underlay) + Service layer (Overlay)
We segment the network
data and applications are made available exactly where they are needed
We modularise the network
simple modules that are standardised and reproducible
We look at everything from an application perspective
– network and connectivity
– security and isolation
– SDN from an application and server perspective
– monitoring und visibility
– reproducibility and reliability
Why building blocks?
- They are standards that can be reproduced.
- They have already been implemented, tested and verified in customer environments.
- They can be individually selected and compiled.
- They can be supported since they comply with the vendors’ specifications and best current practices, and both Xantaro and the vendor are familiar with them.
- All optimisation and continuous development efforts profit from collective intelligence and knowledge: Vendor bulletins or issues in customer set-ups can be solved as part of problem management, and the solution proactively benefits all incarnations of the set-up.
IP Fabric: THE CENTRAL BUILDING BLOCK IN NEXT-GEN DATA CENTERS
The demands that data centers must meet are increasing: Availability, stability and flexibility are a must for all digital transformation processes. Traditional data centers, however, only meet these requirements to a very limited extent: Based on enormous Layer 2 domains and/or Spanning Tree, they are not designed for today’s level of scaling, and their management is time-consuming.
Next-generation data centers are the solution. Their central building block is an IP fabric architecture, based on which services and applications can be produced and implemented without any obstructions and at full speed – even in case of a globally distributed corporate IT.
▷ tried-and-tested routing protocols
Load Balancing (ECMP), i.e. simultaneous use of all existing path
path or system failures are automatically detected and rectified
▷ simple scaling thanks to
the addition of further connections at all layers
addition of further Spine nodes
easy addition of racks at the leaf layer
▷ reduced blast radius
utilisation of compact 1 rack unit systems for Spine and Leaf
▷ open standards
interoperability between vendors and platforms etc. thanks to standardised routing protocols
EVPN / VXLAN, Controller-based or IP – ABSTRACTION OF PAYLOAD DATA
Communication between end points and applications is not implemented directly in the underlay network of the Next-Generation Data Center: The payload is transported through the network using overlay networks.
Depending on the requirements of an application and the infrastructure in place, several different options or a combination thereof can be used.
▷ EVPN / VXLAN
Continued use of all existing Layer-2-based applications and services while simultaneously utilizing all IP fabric benefits
- uses a distributed EVPN control plane and prevents Layer 2 flood-and-learn, i.e.:
– Mac addresses are learned at first occurring and from then known everywhere
– Layer 2 segments are only delivered/held available where they are needed
- open standards enable the cross-vendor utilisation of the EVPN control plane
- VXLAN data plane to transport Layer 2 frames over IP underlay networks; can be used on any Ethernet/IP structure
- overlay termination directly on the Leaf switch enables the simple connection of physical servers and appliances
- layer 2 domains and individual services/customers/tenants can be separated or isolated
- services can be mapped analogously to IP/MPLS L3 VPN to operate L3 services without L2 dependencies; the same service framework can be used for L2 and L3 services
Virtualisation platforms such as VMware NSX, Nuage, OpenStack etc. instantiate their own overlay
- direct, dynamic communication between VMs and/or containers
- prevents flood-and-learn, since the virtualisation solution is familiar with the VMs that are active on the network and their MAC addresses and informs other virtualisation hosts
- a centralised control plane
- networks or tenants within the VM platform and the overlay can be separated
- physical servers and services are integrated based on various gateway solutions (OVSDB, dedicated gateway servers etc.)
- transition to EVPN/VXLAN-based solutions etc. (exchange or communication between the central, controller-based control plane and an existing EVPN/VXLAN control plane, where applicable)
▷ Native IP
The applications communicate directly on the basis of L3 without using L2 services
- use of all IP fabric advantages
- networks or tenants can be separated directly in the IP fabric through Virtual Routing & Forwarding (VRF)
- no legacy dependencies for applications or services
A Distributed Security Concept: Security for Next-Gen Data Centers
Traditional data centers usually rely on large central firewalls for security. Next-Generation Data Centers, however, are protected by a distributed security concept.
- distributed security via Network Function Virtualisation (NFV)
- selective rejection of the first flows and direct delivery of subsequent packets
- in hybrid set-ups, critical applications continue to be routed through central firewalls; security for high-volume traffic is distributed or handled directly on the host
Data Center and/or Cloud Interconnect – individual networking
Modern data centers bridge the gap between core business and digital transformation: They are hubs linking different locations such as company headquarters, branch offices, production sites or outsourced company IT systems.
Any other data centers, regardless of whether they are locally or globally distributed, must be connected, too. Other possible scenarios include private clouds or the connection of public clouds.
All modern data centers must therefore be connected according to their specific requirements and using different transport structures.
▷ Data Center Interconnect (DCI)
to link several globally distributed data centers
- direct optical interconnects for distances from 10 to several hundred km and parallel use of several 100G on the same fibre infrastructure
- L2 links/wavelengths/fibre leased from third parties
- existing IP/MPLS networks for L2 stretching (VPLS, EVPN/MPLS)
- existing IP networks (native or IP/MPLS L3 VPN)
Depending on the requirements, this also allows for the operation of self-sufficient DC networks with independent control planes per site and an overarching network with uniformly distributed or centralised control planes. In addition, further locations can easily be connected at any time.
▷ Connecting Cloud-Services
- for dynamic workload outsourcing, e.g. using AWS/Azure
- attached to the existing data center infrastructure, similarly to Office365 or Salesforce
USE CASE: The data center within the company network
Find out when to rely on a Next-Generation Data Center!
Today’s data centers need to connect different locations (HQ, branch offices, production sites), several data centers or rather a distributed corporate IT, or must interconnect with Cloud Service Providers.
Within this scenario, the data center is the pivotal point that ensures your production/service business runs smoothly at a global level..
Digital transformation and the modern IT applications required place increasing demands on the capacity and performance of data centers. As the heart of the network, the DC infrastructure must guarantee flexibility at the same time as a maximum of availability and stability in order to reliably map business processes.
In reality, however, this is in stark contrast to the technical debt accumulated over the past decades:
- historically grown, rigid and entrenched structures (at technical and organisational level)
- hierarchical networks designed and built according to textbooks of the 1990s
- over 50% unused capacity in the network owing to the use of outdated protocols
- instability and fragility due to scales that were never intended for set-ups this size
- a “no touch” policy, as nobody knows exactly what will happen if something is changed on the network
Traditional DC architectures are now increasingly reaching their limits:
- insufficient flexibility and agility to provide modern applications and processes
- no independent functionality for cross-site applications
- maximum capacity used in terms of bandwidth and physical ports
- error-prone at network layer as a result of the requirements
- security concerns with regard to cloud services and data processed there
Compliance / Date Protection
- Should all “core business” remain in-house as a general rule? How much can be outsourced?
- Is the company data secure within the cloud?
Scaling and bandwidth
- sufficient capacity in the long run
- future growth and new innovative services
- physical limitations (e.g. port density)
Use of modern IT applications
- automated application designs
- container architecture
- global use of Virtual Network Functions (VNF, NFV)
- migrating or outsourcing services is often more complex than expected
- basic and continued training required for all responsible teams
Do you need to keep all IT and assets in-house?
- simultaneously expand resources for new IT applications
- obtain access to hi-tech and cloud technology
- OPEX vs. CAPEX perspective on infrastructure
- lack of transparent and clearly calculable costs
Operating the modern Data Center:
We also support after implementation!
With Yukon, Xantaro provides customers with a specially developed, vendor-independent service engine for cross-technology network and service management.
As a central platform, Yukon collects relevant information from the systems of a network, normalizes the data and puts them within various applications in context. The result:
- Transparency in the network and assets
- Minimized operational complexity and costs
- faster implementation of new services
The use cases are diverse – inventory management or troubleshooting are just two examples of how Yukon is used.
Bug fixes, security issues, new features and systems or rather their interoperability – changes in the network have various reasons. Experience shows that intensive, early and regular tests are crucial since the later the impact, the higher the costs in terms of time, money and reputation.
Our answer to testing of complex scenarios:
XFAST – Xantaro Fully Automated Software Testing
In order to meet the ever shorter innovation cycles and test times, Xantaro offers the XFAST service that significantly reduces test effort, maximizes the quality of test results and minimizes risks when implementing new operating systems into the productive environment! The benefits:
- kundenspezifische Test-Setups
- automatisierte und reproduzierbare Testabläufe
- Durchführung vor Ort oder in unserem XT3Lab
- umfassender Ergebnis-Report
With the services around know-how transfer and training, Xantaro enables customers for the operation of the individual data center solution.
Within know-how transfer, Xantaro offers specially developed training courses – always taking into account the latest technologies and individual developments – to train the customer’s employees. These include:
- Standard training on established solutions
- tailored technology training according to custom requirements
- know-how transfer within individual workshops either “Training-on-the-Job”
The reliable data center operation ensures the availability of business-critical services – another reason why we test all solutions intensively before use and standardize them in modules. This is how we ensure that our Next-Gen Data Center solutions improve stability and increase reliability. The modern architecture also ensures reduced error domains through the logical abstraction of the L2 domains on an L3 underlay, which significantly simplifies troubleshooting.
Furthermore, Xantaro offers with XCare variable maintenance services through to proactive services to ensure reliable operation of the individual data center architecture.
XTAC – the Xantaro Technical Assistance Center – is the central interface towards all our technology vendors. The team consists of dedicated and experienced engineers specialised on the support of complex multi-vendor environments. The benefit: They know the vendors’ escalation processes just as well as the technologies of our data center solutions and are specialised in escalating and resolving faults in a solution-oriented manner.