IHS-Infonetics released excerpts from its IHS Infonetics NFV Hardware, Software, and Services report, which forecasts the global network functions virtualization (NFV) hardware, software and services market to reach $11.6 billion in 2019, up from $2.3 billion in 2015.
NFV MARKET HIGHLIGHTS:
- Service providers are still early in the long-term, 10- to 15-year transformation to virtualized networks
- Revenue from outsourced services for NFV projects is projected to grow at a 71% compound annual growth rate (CAGR) from 2014 to 2019
- Revenue from software-only video content delivery network (CDN) functions for managing and distributing data is forecast by IHS to grow 30-fold from 2015 to 2019
“NFV represents operators’ shift from a hardware focus to software focus, and our forecasts show this. We believe NFV software will comprise over 80 percent of the $11.6 billion total NFV revenue in 2019,” said Michael Howard, senior research director for carrier networks at IHS. “The software is always a much larger investment than the server, storage and switch hardware, representing about $4 of every $5 spent on NFV,” Howard said.
NFV REPORT SYNOPSIS:
The 2015 IHS Infonetics NFV Hardware, Software, and Services market research report tracks outsourced services for network functions virtualization (NFV) projects as well as service provider NFV hardware, including NFV infrastructure (NFVI) servers, storage and switches; and NFV software split out by service management and orchestration (NFV MANO) software and virtual network function (VNF) software, including virtual routers (vRouters) and the software-only functions of mobile core and EPC, IMS, PCRF and DPI, security, video content delivery networks (CDN), and other VNF software. The research service provides worldwide and regional market size, forecasts through 2019, in-depth analysis and trends.
Watch analyst Michael Howard’s July 2015 webinar, SDN and NFV: Accelerating PoCs to Live Commercial Deployment, an event detailing how operators can validate VNFs, NFVI and network services to speed up commercial deployments. Log in to view:
Alan’s DIssenting View of NFV:
There’ve been many times in the last three or four decades where network operators were wildly enthusiastic about a new technology, which never really gained marrket traction. In the mid 1980’s, ISDN was to replace the PSTN phone system and usher in new world of data communicaitons. In the early 1990s it was SMDS (Switched Multi-Megabit Data Service) that never really saw the light of day. In the mid 1990s, ATM was going to take over the world. 10 years later, many operators including SPRINT believed that WiMAX would be the 4G technology of choice. Didn’t happen!
Now we have major telecom carriers all excited over NFV with many forecasts of huge growth of that market. We don’t think that will happen in the next five years or maybe ever. Here’s what’s missing from NFV:
- Implementable standards for exposed physical interfaces and APIs for virtual appliances1 implemented as software running on a generic/commodity compute server.
- Procedures and protocols for virtual appliances to communicate with physical box (legacy, non-virtual) appliances already deployed in carrier networks (otherwise known as backward compatibility with the installed base). An ETSI NFV Use Cases document states in section 7.4. Coexistance of Virtualized and Non-Virtualized Network Functions: “The communications with virtual network functions shall be based on standardized interfaces.” Yet those interfaces haven’t been standardized yet.
- New security methods and procedures to isolate and quarantine compromised virtual appliances so that all appliances running on the same compute server are not locked down.
- Hardware assists, possibly NICs for compute servers, to guarantee latency and throughput of virtual appliances running on commodity compute servers.
- Standardized Management & Orchestration (MANO) software, including “service chaining,” which will work with different virtual appliances from multiple software vendors. Standardized software interfaces (like APIs) will be required for this, perhaps as part of an open source MANO software package (from OPNFV?)
- Element Management Systems (EMSs) that are integrated into MANO. For decades, physical box network vendors provided their own proprietary EMSs which configured, managed, monitored and controlled their equipment. EMSs need to be converted to software modules within MANO or accessible to MANO via software interfaces.
- The ability of MANO or other software to manage, control, schedule appliances/services, etc from BOTH virtual appliances and physical box appliances. Note that both types of appliances will be used in telco central offices/data centers for many years.
- Procedures for testing, monitoring, OA&M, fault isolation, repair & restoration, etc are urgently needed.
- Open Source NFV software, perhaps from the OPNFV consortium. IMHO, this is the best hope for NFV being a commercial success/real market.
- There’s a lot of hype about virtualizing the LTE Evolved Packet Core (EPC) via NFV but that is nonsense because operators are not going to put in a new infrastructure after recently spending lot of money to install EPC equipment and software. Further, there are no standards for vEPC. Yet we see many blog posts/article that say the time is now for vEPC. Here’s one of many NFV -vEPC hype to Pluto blog posts:
Note 1. Examples of virtual appliances include: session border controllers, load balancers, deep packet inspection agents, firewalls, intrusion detection devices, and WAN accelerators.
Here’s an interesting quote from a Spirent white paper on NFV:
“Traditionally, the burden of validating the core functions is shouldered by the network equipment (box) vendors. Today’s NFV landscape shifts some of this burden to operators deploying NFV topologies while not absolving the need for network vendors to validate NFV software within different hardware (compute platforms, physical switching tiers) and software infrastructures (hypervisor, holistic cloud stacks public clouds, etc).”
Tom Nolle on Fixing NFV:
In a New IP blog post, Tom Nolle wrote:
“What’s needed in NFV is something like “NFVI plug-and-play,” meaning that any hardware that can be used to provide hosting or connectivity for virtual functions should be capable of being plugged into MANO and supporting deployment and management.”
Nolle goes on to elaborate on fixes for CAPEX, OPEX and business case. He concludes with this remark: