Archive

Uncategorized

A key promise of SDN technologies is the potential for a vibrant market for “applications” that leverage SDN protocols to address physical network resources. It is SDN’s ability to decouple network logic and policies from the underlying network equipment that allow for a more programmable network. The way to create SDN applications is to give broad population of independent software vendors (ISVs) and other developers the training and tools to leverage open SDN APIs. It is these applications that will deliver the value of SDN to the IT community by allowing greater levels of flexibility, innovation, and control in the network.

The “traditional” network market has struggled to create a broad ecosystem of network-aware application software. ISVs have been challenged by a large number of proprietary network protocols. Several network vendors have promoted efforts to create an ISV community around their specific network environments – most notably Cisco and Juniper- with relatively limited success.

The introduction of OpenFlow brings an open source strategy (e.g. Linux) to the networking market. OpenFlow is offered by a large number of incumbent network vendors including Cisco, HP, Juniper, Brocade, IBM, Dell, and NEC and a variety of start-ups, including Big Switch. There are also a number of open source options for OpenFlow. OpenFlow technology has been widely seeded in leading universities. In order for OpenFlow to achieve broad market acceptance, it must enable a sizeable community of network ISVs.

There are a number of SDN offerings that are not OpenFlow centric that also hope to enable a broad network ISV community. Cisco has announced its Open Network Environment (ONE) with a common set of APIs across its broad product line. VMware with its Nicira acquisition is also creating an SDN application community. Other interesting SDN start-ups with the potential to enable SDN applications include ADARA, LineRate, and Midokura.

2013 will see further standardization of south-bound and north-bound SDN interfaces – this will allow leading-edge ISVs the ability to develop more SDN specific applications. The OpenFlow ecosystem including R&D centers, universities, open-source, and vendor community will see further SDN application development. And, Cisco’s release of open API’s common to all its switch/router platforms (ONE) will enable ISVs and its broad channel community to develop SDN applications.

Development of a broad ISV community to leverage SDN technologies to tie applications to the underlying physical network is critical to the long term success of SDN. Enterprise IT customers want to buy SDN “solutions” that bring specific business value and a wealth of SDN applications will be the catalyst to broad SDN adoption in the business community.

Advertisements

 

Much has been written about the impact of software defined networks (SDNs) in the data center and enterprise network.  The key question that all IT professionals are asking about SDN is what is the business value to their organization in adopting SDN technologies?  SDN technology is new – and therefore time consuming to learn and potentially disruptive to installed networks.  From the IT professional’s viewpoint, the business value of SDNs must outweigh the “cost” (in dollars or time) of SDN implementation.

The business value proposition for SDN comes in a 3 main categories:

  • Lower CAPEX
  • Lower OPEX
  • Network Agility

Lower CAPEX is pretty straightforward.  SDN offers the potential to lower the costs of relatively expensive network gear including Ethernet switches (Layer2), routers (Layer 3) and Layer 4-7 products, such as server load balancers, WAN optimization, firewalls, and IP VPNs.  This CAPEX advantage is especially interesting in green field hyperscale data centers (e.g. Google).

Lower OPEX is a leading target due to the costs to IT organizations of network professionals needed to configure, manage, and support the full Layer 2-7 network stack.  If SDNs can make network operations more self-serve and more automated then this is clear win in terms of business value.

Network Agility refers to the potential for SDN to offer significant improvements on how networks support IT operations.   The classic data center example is reducing the time it takes to provision the network and security services around a new workload or VM.   Another application is the ability to segment (or slice) the network to provide highly secure, multi-tenant services (either public cloud or enterprise).  For Those organizations that rely on their network as a key part of IT (e.g. financial services, high distributed retail organizations, and education), SDNs ability to improve network agility will be additive to business value.

IT professionals looking to implement SDN technologies in their organization should carefully evaluate the business value benefits SDN will bring to their networks in the specific categories of CAPEX, OPEX, and network agility.

COTS (Commercial off the shelf) is short hand for industry standard servers (e.g. Intel), operating systems (e.g. Linux or Unix), and middleware.

The enterprise network and telecom industries have been living in a world of purpose built systems for 30+ years.  It is standard for Ericsson, Cisco, or Juniper to design their own ASIC chips, develop proprietary operating systems (e.g. Cisco IOS), and integrate it all into highly optimized systems (e.g. base stations, routers, and ethernet switches).

Changes in the telecom world have resulted in some COTS adoption as Huawei, Alcatel, and NSN have moved some of their development efforts to commercial servers or ATCA (a blade server standard for high reliable telecom systems).  However, my recent discussions with a range of enterprise and telecom network suppliers indicates a relatively slow migration to COTS hardware and a continued dependence on internally developed middleware.

In the long run, migration to COTS is inevitable given the power of Moore’s Law and proven COTS benefits, namely:

  • Faster design cycles
  • Less expensive hardware
  • Lower maintenance costs
  • Benefits of server virtualization

Networking vendors are increasingly finding their value-add is coming from highly specialized software.   And given financial pressures most networking suppliers need to reduce the time and engineering costs of their hardware, OS, and middleware developments.

In networking and telecom change can (and usually does) take a long time – systems are designed for 5+ year cycles.  Migration to COTS generally occurs when a new equipment design cycle is required (e.g. 4G, 100GB) or in response to specific transformative industry trends (e.g. Software Defined Networking or SDN).

There is significant market opportunity for IT players with targeted programs to deliver COTS to the network/telecom industry.   For server hardware only, the telecom industry represents 8-10% of the total server market.  ATCA represents a $1B+ market opportunity (internal and external development) for COTS.   Additional market value is available in software (OS, middleware, and applications).  Key IT players here are IBM, HP, Intel/Windriver, Oracle, and Radisys.

The key questions are which parts of the telecom industry are ripe for COTS, for example base stations represent a very large opportunity with little to no COTS penetration today.   The price and margin challenges now impacting the large network equipment suppliers will work to accelerate COTS – see the recent financial results for Ericsson, NSN, and Alcatel.  Over time, adoption of SDN will bring the very large Ethernet switch and router markets into COTS scope.

Overview

In mid -June 2012 at its large customer event (Cisco Live), Cisco unveiled a key  SDN/OpenFlow announcement- see http://www.networkworld.com/news/2012/061312-cisco-sdn-260138.html for the announcement overview.   Cisco ONE or One PK is actually a series of products that roll out over the next 6 months to enable customers to better program their networks.

Benefits

Cisco has made a comprehensive announcement that will allow its customers to start the migration to SDN with or without OpenFlow.  Programmability can take place at various levels of the network stack from transport through application (management) layers.  This is very much an evolution (not revolution) that will appeal to the very large Cisco installed base.   The announcement of ONE PK gives Cisco a clear SDN strategy for its switching and routing products.

Opportunities/Challenges

Cisco is a large company with dozens of products that directly or indirectly relate to its SDN strategy.  It will need to carefully coordinate engineering efforts to sync its release trains – and avoid confusing customers.

Cisco customers and partners will require significant education as to Cisco SDN and OpenFlow offerings.  One way to address this is for Cisco to showcase internal and external development releases that truly illustrate the benefits of Cisco ONE.  Cisco needs to build an “ISV” community where customers and partners can easily see and leverage development code, and APIs that utilize ONE PK.  Cisco should look at various open source efforts for best practices.

Cisco needs to integrate its SDN strategy with the rest of Cisco – large parts of Cisco relate to this strategy including:

  • Its Data Center team (server, storage, etc) – hyperscale data center users will be leading edge SDN adopters
  • Unified Communications – UC  (voice, video, WebEx, etc) are a series of applications that can benefit from improved network programmability
  • Cisco Prime Network Management – Cisco has not explicitly tied Prime to its ONE announcement, SDN is all about making the network easier to configure, change, and manage
  • Cisco Security – Changes in network architectures (e.g. SDN) will require new security paradigms, Cisco security products should be linked closely with its SDN offers
  • Service Provider – SDN has the potential to significantly impact telecom infrastructure (not just in the data center) – SDN should be related to Cisco’s ability to improve CSP operational agility

Recommendations

SDN and OpenFlow have broad implications for the future of the network industry in terms of products, market share, and margin.  Cisco should leverage the “buzz” around SDN in general and OpenFlow in particular to articulate its broad vision for the future of the networked enterprise (e.g. what is its vision for the enterprise in the year 2020).  This would take the form of a broad architectural vision with multiple phases – for those who remember SONA take 2.   This will help the broad range of Cisco internal stakeholders, partners, and customers to understand the future of the intelligent network and how the Cisco SDN’s strategy will help to make networks more flexible, easier to manage, secure, and programmable.

My 25+ years of experience in analyzing the IT, enterprise network, telecom, UC, and security markets causes me to reflect on some of the hot topics/questions that I believe will impact the intelligent network over the next five years.  I firmly believe that it will be an Intelligent Network– one that will require scalability, reliability, …

Read More

Leading Themes for Intelligent Networks

Having 25 years plus of experience in analyzing the IT, enterprise network, telecom, UC, and security markets causes me to reflect on some of the hot topics/questions that will impact the intelligent network over the next five years.  I firmly believe that it will be an Intelligent Network– one that will require scalability, reliability, security, quality of service, programmability, etc.