ACG Research

ACG Research
We focus on the Why before the What
Showing posts with label Eve Griliches. Show all posts
Showing posts with label Eve Griliches. Show all posts

Monday, March 18, 2013

The Economics of Cisco’s nLight Multilayer Control Plane Architecture


Networks are becoming more difficult to plan and optimize because of high traffic growth, volatile traffic patterns, changes in network architecture and more stringent service level agreements. Because the IP and optical control planes are not integrated, service providers’ networks are always over-engineered and underutilized with 50–75 percent of the network capacity deployed for network resiliency, not for passing traffic. A more economically efficient and effective network resiliency solution is needed.

Cisco has addressed this problem with nLight: a multilayer routing and optimization architecture that focuses on IP and DWDM integration, increasing network agility and flexibility while improving network utilization. ACG Research compared the nLight approach to protection and restoration to the widely used 1+1 optical protection scheme. Click here to download the TCO.


For more information about ACG Research's business case analysis services, contact sales@acgresearch.net.

mkennedy@acgresearch.net
www.acgresearch

Juniper Networks Delivers the PTX3000 with 100G Integrated OTN


The proliferation of 100G connectivity in core networks is now spreading to the regional and metro/access areas. The demands in the metro access area are, in fact, similar for scale, connectivity and latency but actually up the bar when it comes to footprint, power utilization and lower BTUs per Gpbs.

Juniper Networks introduced the PTX5000 at OFC in 2011 and shook up the router landscape with the introduction of the first core-optimized  product on the market that delivers a simple architecture, which both content and service providers can leverage. 

The company has now introduced the PTX3000, which will play an important role in the metro core as Juniper Networks is now delivering the power of the Supercore in a transport package.





Eve Griliches
egriliches@acgresearch.net 
www.acgresearch
 



Tuesday, March 5, 2013

Emergence of Optical Networking in the Data Center


The networking industry is on the cusp of a major data center transformation where optical networking technology will play a pivotal role in shaping the next data center architecture. Fixed network topologies are adequate when data traffic is relatively predictable; however, data centers, today, are experiencing dynamic traffic flows that are unpredictable and require unprecedented amounts of any-to-any connectivity. In addition, server virtualization, workload mobility, and big data applications are changing traffic patterns from north-south to predominantly east-west, which places strains on typical spine/leaf networks. 

With workload mobility, virtualization, and cloud computing all vastly changing the data center, introducing the benefits of optical networking technologies into products that can manage network capacity, traffic flows, and scale to support thousands of server racks is bound to hit a sweet spot in the data center. Which company is incorporating multiple optical technologies into the data center and leveraging SDN? Click here to download the Market Impact.




Eve Griliches
egriliches@acgresearch.net 
www.acgresearch
 


Tuesday, February 19, 2013

100G Port Counts, It's Time to Take a Step Back

The current market for 100G ports is hotly contested and competitive, suggesting increasing disparity in the market. 

100G port shipments are simply not lining up with deployments or revenues. ACG Research tracks 100G ports but does not report on them for this very reason. Let's take a look at some of the factors influencing the market:
  • Route distance: In some cases 100G will require regeneration, some will not. Most 100G deployments are layered on top of 10G routes so that regeneration sites are already established, but with coherent technology, longer distances are clearly achievable. If a service provider deploys a longer route, there will be fewer ports. If the provider has to regenerate, the port count will double. Thus, long-distance products can have lower port market share and are “penalized” simply by having better technology. In other cases, the routes are shorter, and more regeneration is an accepted fact. Either way, the numbers get skewed. 
  • 100G ports are often split between metro and long-haul deployments; yet, the definition for metro can be widely assumed to be anywhere from 300km up to 700km. 100G total port counts may be recorded, but disparity of metro definitions by vendors and analysts adds to the confusion.
  • Some shipments are recording nonrevenue field trials, spares and free products as port shipments (not a standard process) and this presents a challenge to defining what constitutes true port counts.
  • As we are in the first year of 100G shipments, there have been the usual technical issues with any new and significantly higher bit rate. This often results in additional sparing or replacement parts, which are added to the total shipped port counts.
  • In some cases, ports reported in previous quarters have been revised upwards.

We know that 100G pricing has been extremely volatile this past year. To determine an average selling price of a 100G port becomes even more problematic when revenues cannot be tied to shipments. 

Success metrics at this time should not be based on port count. Inherently, there are flaws in the current market that are distorting market share. Moving forward, we expect this to change as the market matures. As we see the price of 100G stabilize and the technology maturing, the situation will improve. 

The following are some guidelines if you are looking for overall value and sustainability in the market:
  • Review other success factors in the segment. Are revenues trending up in a difficult market?
  • Which vendors are service providers selecting? Are they announcing them?
  • Request names of customers deploying 100G. Most use two vendors (some exceptions) for 100G.
  • Examine the deployment roll-outs. How large are they?  One route does not a network make.

Note: we target no specific vendor or analyst with this report. These issues will eventually resolve themselves. We merely want to clear the air now until there is better visibility.

For more information about ACG Research's Packet Optical Transport click here or contact  sales@acgresearch.net or contact Eve at +1-408-200-0967.

Eve Griliches
egriliches@acgresearch.net 
www.acgresearch
 


Worldwide Optical Market Posts Increases


The worldwide optical market increased 8.9% quarter-over-quarter and decreased 5.2% year–over-year. Last quarter, we projected that providers would post revenue of half a billion dollars below 2011 total spend, which is exactly what happened, because 1) new equipment and 100G revenues were down ~30% in 2012 compared to shipment revenues in 2011, and 2) major Tier 1s have excess capacity on their networks which is slowing their migration to 100G or causing them to move at a slower pace to upgrade capacity only where it is really needed. 

Once again content providers dominated the market; deployments were up 50–70% although the revenue paid to vendors was down 30%. Positive points show accelerating revenues for almost all vendors not focused on Tier 1s. Reported profit margins were up for the same group. We expect this trend to continue into 2013.


egriliches@acgresearch.net 
www.acgresearch
 


Thursday, February 14, 2013

Next-Generation Inter-Data Center Cloud Infrastructure


Cloud services and increased machine-to-machine (M2M) connectivity have contributed to the explosion of inter-data center traffic. 

Cloud computing has redefined the way applications run on the network, exposing the underlying limitations of the networks. To keep up, content providers have changed the paradigm and built their own large capacity networks. However, operational complexities and cost continue to be a challenge as bandwidth and applications grow at a rapid pace on the network.

The market, therefore, needs a new solution specifically designed to address this growing problem. Next-generation cloud infrastructure platforms must meet massive capacity and scaling requirements. Content and service providers need converged solutions that include optical and LSR functionality (see http://www.acgresearch.net/knowledge-insights/reports-whitepapers/next-generation-lsr.aspx), especially in the metro regional networks in order to lower the cost per bit of transport and routing. Adding application awareness to a converged solution will also enable content and service providers to meet the performance and scale demands while increasing control and enabling service innovation, all with significantly improved economic efficiencies. 




 
egriliches@acgresearch.net 
www.acgresearch
 



Tuesday, January 15, 2013

Join ACG Research at the SDN Precon at ITEXPO Miami


Software-defined networking is considered by many to be one of the most important and potentially disruptive developments in networking since the rise of the Internet. To date, networks and connectivity have come first, and applications have followed. SDN turns this model on its head, approaching networking from an apps-first standpoint. In this SDN preconference at ITEXPO Miami, we (including ACG Research’s Eve Griliches, optical analyst) will:
  • Define SDN
  • Look at the technologies and components involved in SDN
  • Review the use cases of SDN
  • Identify the players involved in moving SDN forward
  • Discuss what this all means for networking as we know it

For more information about this event, visit: http://www.sdnzone.com/conference/

For more information about Eve Griliches, click here.




egriliches@acgresearch.net 
www.acgresearch
 

Tuesday, December 18, 2012

Evolving IP and Optical Architectures

ACG Research's whitepaper "Evolving IP and Optical Architectures" discusses the challenges service providers  are facing and explores the pros and cons of next-generation networks.

The inflection point of 100G being deployed in the network creates the perfect opportunity to re-evaluate and optimize your network, click (http://www.juniper.net/us/en/dm/acg-supercore-wp/) here for your free white paper on how today’s network architectures are evolving and how operators will benefit by moving to new and more cost-efficient network architectures.

For more information about Eve Griliches, click here.




egriliches@acgresearch.net 
www.acgresearch
 




Wednesday, December 12, 2012

SDN Business Models: Which Will Succeed?

As I see more and more companies enter the software-defined network (SDN) market, my hardware-centric mind struggles to understand the business models of many of these software offerings. After talking to industry experts I have come up with four basic models:

1. Embedded Model. This model is hardware based. System vendors will likely throw in the controller and open APIs for free and price the hardware accordingly; the hardware covers the cost of the software development. This technically obscures the software solution within the hardware price. Benefits of this business approach is that the hardware stays in the network a long time, and additional add-on software features and/or applications and upgrades can be sold at increasing profits. This model will be used by key networking infrastructure vendors as well as SDN start-ups that have hardware-based solutions wrapped together with the controller and applications. This model is likely to be the most successful in the long run, because end users are leaning toward a full software and hardware package versus piecing the components together themselves.


2. Cloud-Based or Application Hosting Model. This model is per virtual machine (VM) and is time based, similar to the Amazon usage-based billing system. Use a number of VMs within a certain period, and the price scales with the usage. There is no tie to specific hardware. It may sound like a low-cost entry approach, but this is likely to be one of the most expensive models because the complexity to evaluate the year-end cost is quite difficult. Additionally, it is likely too expensive for the typical enterprise to embrace given large-scale virtualization deployments are in the very early stages. 

3. Subscription Based. The subscription-based model simplifies things a bit for the customer and enables either a quarterly or annual fee to be attached to the controller and applications purchased. The controller can be separate from the applications or bundled, enabling multiple offerings that are a single price. Upgrades can be priced by quarter or year, which enables up-selling of new applications in a very straightforward format.  

4. Capitalized Model. The capitalized model is where the customer purchases the controller (and/or applications) for a one-time price and capitalizes it over time. This is where you find out where the purchasing power is in the organization. Often the networking team will submit a VM or subscription-based proposal, and the CFO will insist on a one-time price instead.

Several of these business models, in theory, are compelling, but have become extremely complex for the end user to evaluate, specifically, determining what the end cost of the network will be. End users are telling us that the simpler the business model, the more likely the sale, given the CFO influence. Thus, the capitalized model, in many cases, is actually the most popular right now, despite what the market is telling you.

For more information about Eve Griliches, click here.

For more information about ACG Research's SDN services contact sales@acgresearch.net.



egriliches@acgresearch.net 
www.acgresearch
 


Wednesday, December 5, 2012

ECI Delivers on Native Packet Transport

Packet optical transport products were introduced to the market to handle the ever increasing packet traffic in an affordable manner by lowering the cost of transported bit.  However, simple and low-cost packet-based equipment may lack carrier-grade transport attributes. Adding attributes such as protection, redundancy and resiliency to packet-based products has proven to be more difficult than expected, often resulting in higher capital costs related to addressing these requirements.

Managing large-scale networks also has proven to be more expensive than anticipated and operators have started to conduct thorough total cost of ownership (TCO) analysis for their networks, which provide evidence that operation expense (OpEx) is an important factor that impacts overall considerations. In addition, although Ethernet services are growing rapidly, it has become clear that mixed traffic is still running over most networks, and this situation is not expected to change in the foreseeable future. Consequently, operators have found that the overall cost of a packet-based transport network is more expensive than initially anticipated.

See ACG's Product Impact on ECI’s Native Packet Transport (NPT). 

For Michael Kennedy's TCO on ECI's MPLS-TP NTP Solution, click here.

For more information about Eve Griliches, click here.




egriliches@acgresearch.net 
www.acgresearch
 



Tuesday, November 27, 2012

Multiple Factors Reduce Optical Network Forecast

ACG Research has reduced the Optical Networking forecast for 2012-2017 from a 3.6% five-year CAGR to 2.6%.

    • Economy: The economy continues to affect the market as regions such as the euro zone have not improved, and North America has not built any significant momentum toward a clear or immediate surge in spending. There seems to be few outlets for positive data points.
    • Capital Expenditures: We are not seeing the CapEx flush that some expected in the second half of 2012. While AT&T has reported increased spending for broadband over the next three years, we believe only a portion of this will trickle down to optical expansion requirements.
    • Guidance: Several vendors are guiding flat or are lowering expectations for Q4 (Alcatel-Lucent is an exception). Those heavy in the enterprise have reported corporate trepidation and are slowing their spending because of indecision.
    • Company Size: We are seeing smaller companies doing better than some larger ones primarily because they are well diversified in their customer bases. Most medium or smaller companies are selling to content providers as well as the enterprise, which helps buffer the lower spending by Tier 1s. Those that are largely Tier 1 focused are feeling the most pain.
    • Pricing Pressure: 100 G pricing and general ASPs have fallen more than 50% in the last 5 quarters for some vendors.
    • Product Management Choices: Vendors that have multiple products are not making the difficult product management decisions to move forward and support key programs. Research and development is being spread across multiple platforms, resulting in delayed product and lack of innovation.
    • Consolidation: We expect these factors will finally result in consolidation in the optical market.

                ACG continues to be cautious in our overall estimates for the next five years. The transition from separate IP and optical platforms to converged packet optical-based platforms remains slow as operators weigh the advantages and disadvantages of integrated solutions and the effect of SDN on the overall marketplace.

                For more information about Eve Griliches, click here.

                For more information about ACG Research's SDN services contact sales@acgresearch.net.



                egriliches@acgresearch.net 
                www.acgresearch
                 

                Friday, November 16, 2012

                Optical Market Decreases in Q3 2012


                In addition to economic uncertainty, vendors in the optical market are dealing with more intensive competition and diminishing profit margins as well as seeing their largest customers cut spending and delaying purchases.

                The Worldwide Optical market decreased 3.7% quarter-over-quarter and remained relatively flat year–over-year. We project that 2012 will post revenue of half a billion dollars below 2011 total spend as service providers have informed us that we should not expect a CapEx flush in the last quarter of 2012.

                Bright points in the numbers show BTI continued its upward growth quarter-over-quarter (15%) and year-over-year (45%). Infinera also posted very positively, up 27% quarter-over-quarter and 10% year–over-year. Both ECI and Alcatel-Lucent were down.

                We expect market consolidation because of price erosion. Some vendors will merge or sell off assets or will have to terminate product lines. All options are being explored to bolster cash balances to more normal levels. 




                Eve Griliches
                egriliches@acgresearch.net 
                www.acgresearch
                 

                Thursday, August 16, 2012

                Growth Highlights Q2 Optical Market


                The total optical market was up 25 quarter over quarter and there indicators that some operators have held back their CapEx and may spend in the next two quarters.  

                In 2008 Alcatel-Lucent had 21% market share and was the number 1 vendor in the optical networking market. Today, Huawei and ZTE combined have 40% of the entire optical networking market share. For the first time, Alcatel-Lucent has dropped to number 3 in overall market share (12.6%). ZTE rose to the number 2 position. 

                Spending in the content provider and the enterprise spaces is floating a significant portion of the market, especially for 100G vendors and for smaller vendors that have specific equipment for content delivery or feature sets for enterprises’ low latency networks.


                We are cautious about the market for the remainder of 2012 and 2013. Pricing has taken a serious toll on revenues and although discussions about new architectures to improve margin are still in the discussion phase, we are not entirely sure how much savings will occur in the optical space.

                For more information about Eve Griliches, click here. For more information about ACG Research's optical service click here or contact sales@acgresearch.net.



                egriliches@acgresearch.net 
                www.acgresearch
                 

                Monday, June 18, 2012

                Ciena Financial Analysts' Conference

                Eve Griliches was in NYC last week for Ciena's financial analyst conference. Click here to listen.

                For more information about Eve Griliches, click here. For more information about ACG Research's optical service click here or contact sales@acgresearch.net.



                egriliches@acgresearch.net 
                www.acgresearch
                 

                Switching Architectures and Implications on Network Efficiency


                Operators realize that in order to better monetize their networks they must significantly reduce their time to market delivery for new services while keeping capital and operational expenses under control. These challenges require a newer and more efficient network. Almost every operator is examining changes to their core and metro architecture to address these needs, either by migrating to a new architecture or technology or with an overlay of one. 

                In  "Switching Architectures and Implications on Network Efficiency" we discuss new architectures, why service providers like them, and which architectures best promote network efficiency and flexibility.


                For more information about Eve Griliches, click here.




                egriliches@acgresearch.net 
                www.acgresearch

                Wednesday, May 30, 2012

                Talking with Plexxi, Inc., Part 2

                Eve Griliches, optical analyst for ACG Research, continues her interview with Mat Mathews, vice president of product management; and William R. Koss, vice president, to discuss the data center market, SDN and what Plexxi, Inc., is doing.

                Click here to read Part 1 of the interview.

                Click here to read Part 2 of the Interview.

                For more information about ACG Research's packet optical transport syndicated and consulting service, click here or contact sales@acgresearch.net.



                egriliches@acgresearch.net 
                www.acgresearch

                Monday, May 14, 2012

                Cisco and Shifting Paradigms


                The two gentlemen were already in the elevator when I got in. The third floor button was lit up so I stepped back and relaxed. One of them said "you want the third floor?" I said yes. He remarked "that's amazing." I said "Not really, there was a 25% chance (the hotel had four floors) I'd be on the third floor." Now for you game theorists and math majors you'll know that my percentage was way off and that the probability of events like this is never very intuitive. The first floor should never have counted because no one uses the elevator. So assuming that, we are now one in three or a 33% chance.  But how many folks on the second floor take the elevator? So, what appeared to me to be easy math was actually a significant miscalculation.

                How many of these calculations do we do every day? If the paradigm we live in is transparent, hopefully few. Having the sense to know we need to look at things differently started in that elevator ride and continued during my visit in the Dallas area. We are simply not going to build networks the same way we have done in the past. Talking to some content and cable providers there, it seems that the networks being built by these guys are showing us it can be done differently today and that new paradigms are, indeed, possible. Look around and see how many networking companies are still in the market today, have strong portfolios, and have not fallen into the traps of yesteryear. Very few. Those companies that have continued to shift and adjust, fallen and gotten up, and jumped back to work are still here today. Cisco is one of those companies. And, although I rarely comment on stock value, Cisco is likely worth much more than the probability assigned it today. As paradigms shifted, so have they.  

                Thursday, May 10, 2012

                Talking with Plexxi, Inc.


                I met with Dave Husak, founder and CEO of Plexxi; Mat Mathews, vice president of product management; and William K. Koss, vice president; to discuss the data center market, SDN and what Plexxi, Inc., is doing.




                egriliches@acgresearch.net 
                www.acgresearch

                Thursday, April 5, 2012

                OFC/NFOEC 2012 Wrap Up

                Even before OFC/NFOEC 2012 officially got underway key messages about data center innovation and the critical need for photonic integration for low-cost, low-power interconnects, as well as the need for increased interest and investment in this integrated component area were the buzz at the conference.

                Workshops featured deep dives into data center architecture, and discussions varied from the distributed all-to-all Hadoop approach to thegather-scatter centralized approach.

                In other workshops participants pondered issues such as having to throttle back data rates between servers and potentially turn off processors (sleep mode) to save power or how to determine the right network design that supports fast data flow and also utilizes optical pipe switching at lower power levels.

                Highlights

                OFC/NFOEC 2012 was all about the data center. Inter DC continues to focus on scalable and simple solutions in the long haul and metro networks connecting DCs worldwide. Metro deployments are expected to be in full force in 2013 and 2014, especially if low-cost coherent technology is available.

                Data interconnects are clearly the major source of bottleneck within data centers today. The industry is looking for low-cost optical interconnects to replace copper because of distance and bandwidth limitations. By using silicon photonics, optics can advance the state of art in the computing environment into a faster and more scalable data center. Silicon photonics will likely be the building block for a growing number of computing and networking products. Reduced size, power and high-volume deployment will enable more low-cost options.

                Other discussions centered on the bandwidth gap. With 40 percent average traffic increase, optimistic improvements in physical hardware can help by about 10 percent. The industry needs long-term research to deliver some innovative changes as we push towards Shannon's limit for capacity.

                A Final Note

                The entire ecosystem is strapped for dollars and seems to be operating on much lower margins. But the feeling at the conference was very upbeat. While margins are not great for everyone, it does feel like winter is over and investment is back into start-ups, although slight, and positive outlooks are everywhere.

                Paul Saville, Level 3
                Eve Griliches, ACG Research
                Donn Lee, Facebook
                Bikash Koley, Google

                Thursday, March 15, 2012

                Operational Efficiency of Super-Channels

                Eve Griliches, ACG Research, and Dave Welch, Infinera, discuss the benefits of flexible modulation and OTN integration


                Watch Eve Griliches, packet optical analyst at ACG Research and Dave Welch, cofounder and executive vice president at Infinera, explain, using an example, the operational efficiency of super-channels, the benefits of flexible modulation and OTN integration.

                Infinera has integrated the OTN fabric into the DTN-X, providing significant operational benefit while meeting power, footprint and bandwidth requirements. With bandwidth outpacing chip development, super-channels are the key technology to enable higher bandwidth deployment today.

                Eve and Dave explain why 16-QAM is limited in reach for all vendors and carriers, and what the implementation penalties are. They also discuss the increased spectral efficiency and how it is still a win-win for shorter, highly concentrated routes and how with one line card in the DTN-X, you can configure sub-terrestrial distances and high bandwidth congested routes, which enables a new paradigm for optical networking.