Cisco Blogs


Cisco Blog > Data Center

Federated ACI Fabrics for Dual Data Center Deployments – A Disaster Recovery ACI use-case

Wins, Accomplishments, Fast Action, welcome to the world of Cisco ACI. In this blog, I want to take you closer to the core of ACI excitement. Cisco Insieme Business Unit and Cisco’s premier Partner, World Wide Technology Inc (WWT) have come together in developing an ACI based Business Continuity/Disaster Recovery (BC/DR) solution for the next generation Data Centers. This blog specifically addresses the Disaster Recovery ACI use case implemented in WWT’s Advanced Technology Center (ATC). I will present highlights of how ACI has been implemented at ATC as two fabrics, across two Data Centers with federated controllers implementing an autonomous infrastructure and with replicated tenant configurations that will provide for disaster recovery.

This BC/DR use case couples the storage replication solution by Zert0 on NetApp storage and with a completely integrated and consistent ACI network solution on the primary and secondary sites to enable rapid application bring up on the remote site.  Network and security policies are replicated, compute resources are virtualized and synchronized, and storage is continuously replicated.  This integrated architectural approach addresses one of the major challenges enterprise customers have in deploying BC/DR solutions – aligning the configuration and deployment of network infrastructure in a simple process with the storage and application teams to achieve the Recover Point and Recover Time Objectives.

Network architecture: The ACI based network architecture is comprised of two independent fabrics with L3 connectivity between them. Each data center has a unique IP addressing namespace scheme and connects to the WAN. In the operational model per diagram-1, the “East” Data Center is termed primary and the “West” Data Center termed the backup (disaster recovery). Each Application Policy Infrastructure Controller (Cisco APIC) controller cluster is identified as the primary or secondary instance, and changes, additions or deletions to the application tenants, are replicated from the primary to the backup controller. Application tenant configurations are managed through a special Python module developed by WWT that programmatically synchronizes the two fabrics.

BCDR-1

External WAN connectivity for each Data Center is provided through the common tenant in respective ACI fabrics. By using the common tenant for external connectivity, the network and security administrator can assign the appropriate network configuration policy, security contracts and as well as firewall and load balancing services for the fabrics in each data center. The application (DevOps) teams will reference the common configuration and configure application connectivity for intra- and inter-tanant communications through the Application Network profile (ANP). F5 Gobal Traffic Manager (GTM) allows holistic management of multi-data center application delivery via intelligent DNS.

BCDR-2

This ACI based Disaster recovery solution has several other facets like storage replication, orchestration software (developed in-house by WWT) among other solution components. Please watch the YouTube Video for a demo illustration and the whitepaper for design details.

In closing, some key takeaways. Cisco ACI’s innovative architecture enables enterprise apps to treat the Data center as a dynamic, shared resource pool. This pool of resources is managed through a central controller (Cisco APIC) exposing all configuration and management components through a northbound REST API.  WWT exploits this programmatic interface of ACI to develop business continuity/disaster recovery solutions for customers.

Related Links

https://www.youtube.com/watch?v=XnA1P1aYYYc&list=UU-ZgkyIop9MqOL4QF80UU9Q

www.cisco.com/go/aci

www.cisco.com/go/apic

https://www2.wwt.com/sites/default/files/White%20Paper%20-%20ACI%20and%20Disaster%20Recovery%20Solution.pdf

Tags: , , , , ,

Announcing the new Data Center and Cloud Community!

It’s finally here- the new Data Center and Cloud community framework has launched! We created new content spaces for Compute and Storage, Software Defined Networks, Data Center and Networking, and OpenStack and OpenSource Software.

communitydatacentercloud

Cisco Data Center and Cloud Community Infrastructure

Read More »

Tags: , , , , , , , , , , , ,

Deliver Services at Internet Speed

How quickly can your organization stand-up a new application or deploy new services?  Most customers tell me, “not fast enough!”   I am clearly hearing from them that the new standard expectation across the organization is to receive precise data center resources in “internet time,” easily and definitely on-demand.

But customers are not the only ones affected by these new expectation standards.  Application developers also expect to receive the resources they need to support their efforts within one hour — without a lot of process meetings and repetitive, slow paperwork.  They want what they want, when they need it, which is always now!  Can’t get it now?  Out comes the credit card and they go on a shopping spree to outside resources.

Developers don’t worry about security, governance or quality of service.  If you are in operations, or you’re a C-level executive, you care.  You need to meet compliance guidelines.  So how can you get everyone on the same team, working together so the organization can succeed, the old “win-win-win?”

At CiscoLive Milan in January, we introduced the Cisco ONE Enterprise Cloud Suite. Watch this replay of our live broadcast.

 

Read More »

Tags: , , , , , , , , ,

Now is the Time to Migrate from RISC/UNIX to Cisco UCS

Why is now the time to migrated from your proprietary RISC/UNIX platform to Cisco UCS with Red Hat UCS migrationEnterprise Linux. There are several reasons: Technology transitions are driving new demands on data center and IT infrastructure. These new applications and computing models have standardized on x86 architecture while aging RISC/UNIX infrastructure doesn’t provide the performance or the flexibility required to support the needs of the business. IT departments may also be experiencing the pain of RISC/UNIX maintenance and software licensing costs while their budgets are flat or shrinking. At the same time, overarching uncertainty about RISC/UNIX futures foreshadowed by rapidly declining market share combined with dropped hardware and software support. According to the most recent IDC Q42014 server tracker, RISC revenue has dropped 40% from 2012 to 2014 while Cisco UCS revenue has increased 80%.

Read More »

Active SLAs – New Approach for Multi-sourcing Governance

Last month, I attended the International Association of Outsourcing Professionals (IAOP) Outourcing World Summit in Phoenix, Arizona. I had the chance to glean fresh insights from industry-leading practitioners and get a global view from sourcing executives.

Though there has been an evolution in the outsourcing industry, one thing has not changed: outsourcing is a winning strategy. In the recent study from IAOP & Information Services Group (ISG) Annual State of the Industry 1 presented at the event, it stated that Multi-sourcing as an outsourcing approach increased by 75% from last year’s percentage”. The predictions for 2015 also show continued growth: “51 percent of respondents in the survey said they expect to pursue more outsourcing opportunities during the year”.2

Multi-sourcing - where a customer uses several service providers to provide a single scope of outsourced services, is now the new norm in outsourcing. In today’s organizations, outsourcing is increasing and each additional outsourcing agreement makes it more difficult to holistically manage IT service & support.

A new way of thinking about SLA’s is required to make multi-sourcing effective

Service-level agreement (SLA) management is the process of negotiating, defining and managing the levels of IT service. SLA management is a key challenge that organizations face as part of outsourcing and it is becoming increasingly difficult with multi-sourcing.

Most organizations invest heavily to define SLAs. If you talk to the service tower owners themselves, fewer than half have know what SLA’s are actually in the contract. They also cannot recall all of the specific details that the increasing numbers of the agreements contain. So, there is a lot of work that goes into developing these structures but there is no framework established on how to govern them in an automated way.

As a result, most SLA management is reactive. Periodic reviews between buyer and vendors are based on summary performance data from the prior period. Data from vendors is typically inconsistent and inaccurate. I was speaking with a sourcing executive at a large financial services company a couple of months ago, and she reported that half of the meeting time spent with vendors on performance was used just to agree on what an accurate view of the data was. This hardly aligns to the dynamic nature of business today.

New Paradigm on Thinking about SLA’s

Imagine the ability to manage SLA’s with each of your providers in real-time with one service integration framework. In a workshop we held in New York with a number of large buyers and providers late last year, they highlighted that this would not only move the service management model from reactive to proactive, but would also enable a more collaborative working model – moving away from the “us vs. them” paradigm. In one multi-sourced organization we’ve implemented this model in, it has completely eliminated after-hours escalation calls and dramatically reduced resolution times. With an active SLA capability you can manage all of the multi-sourcing relationships from one “pane of glass” and get accurate service information in real-time, such as:

  • Overview of all your open cases per provider
  • Cases trending towards SLA violation that you can proactively take action on
  • Actual violations that require immediate attention

To be effective, this solution would need to be very easy to consume, showing only what really matters up front. Front-line managers don’t have time to browse hundreds of tickets when only looking for one affecting the SLA. This all sound promising but is it possible to achieve?

Transforming multi-sourcing with active SLA’s

We are pleased to announce the availability of Active SLA Management in the 7.0 release of Cisco ServiceGrid. With a highly innovative user experiences via a centralized dashboard and reporting capability for the connected multi-sourced ecosystem, you will gain real-time visibility & drill-down functionality for events trending toward or violating SLAs to enable a more proactive, end-to-end vendor management capability.

If you would like to know more, feel free to browse:

 

Join the Conversation:

 

Sources:

  1. Annual State of the Industry Jagdish R. Dalal, IAOP
  2. Outsourcing 2015: Changing in a Good Way – Contributor: ISG Research, February 2015

 

Tags: , , , , , ,