EtherealMind

Software Defined & Intent Based Networking

You are here: Home / Archives for Cisco

Musing: FTC Accepts Proposed Consent Order in Broadcom $5.9 Billion Acquisition of Brocade

By Filed Under: Cisco, Musing

  1. Fibrechannel competition is reduced to two suppliers, Cisco & Brocade.
  2. The merged Broadcom/Avago is silicon-first company that manufactures and designs ASICs as a core business
  3. I believe that Cisco outsources some part of its silicon design to third parties.
  4. Cisco outsources the manufacturing of silicon to third parties including Broadcom.
  5. Cisco now buys fibre channel silicon from its only competitor.

Brocade and Cisco are the only two competitors in the worldwide market for fibre channel switches, and Broadcom supplies both companies with ASICs to make fibre channel switches. The complaint alleges that Broadcom’s acquisition of Brocade could harm worldwide competition in the fibre channel switch market because as Cisco’s supplier, Broadcom has extensive access to Cisco’s competitively sensitive confidential information.

Probably because no one actually cares about Fibrechannel I guess. Margins are good, customers are suckers and its time to make stupid money from stupid customers.

FTC Accepts Proposed Consent Order in Broadcom Limited’s $5.9 Billion Acquisition of Brocade Communications Systems, Inc. | Federal Trade Commission : https://www.ftc.gov/news-events/press-releases/2017/07/ftc-accepts-proposed-consent-order-broadcom-limiteds-59-billion

Cisco’s Has a New CEO as John Chambers stays on.

By Filed Under: Blog, Cisco

Some quick musings on John Chambers finally stepping down as CEO of Cisco.

  1. Chambers is still Executive Chairman and Chairman of the Board. He isn’t going far and not much will really change in the next 12 months.
  2. A personal thanks for making my unbelievable career possible.
  3. A change at the top of Cisco seen as long overdue by most industry watchers.
  4. The EMCworld conference is happening in Las Vegas. Cue your own speculation on that. Remember that the global Cisco Partner conference was last week.
  5. Chuck Robbins wasn’t the front-runner. Rob Lloyd and Gary Moore were previously tipped and suggests that something has changed inside the Cisco boardroom.
  6. Chuck Robbins was running sales and operations – you could speculate that the board wants to focus on defending and continuing the current sales position and improving internal efficiency. Moore/Lloyd were positioned as leading the “next generation” of technologies that Cisco that have moved slower than expected.

Cisco press release has some “Why Robbins ?” points:

He most recently served as Cisco’s senior vice president of worldwide operations, leading the company’s global sales and partner team that drives $47B in business for the company. He has helped lead and execute many of the company’s investments and strategy shifts, including building the industry’s most powerful partner program, now worth more than $40B in revenue to the company each year. He was also a key architect of the company’s strategy for the commercial business segment, which grew 8% year-over-year last quarter, and now represents 25% of Cisco’s total business

You could put almost any sort of speculation into this space as to why the previous front-runners weren’t selected. I would take a guess as follows:

  1. The BUs that Chuck runs appear to have been stellar performers.
  2. Cisco cares a LOT about sales in the current market as it is under attack from all sides.
  3. Sales in the Data Centre haven’t been that great compared to overall growth in the segment. UCS has done very well but remains small, ACI has a had slow start. Importantly, VMware’s NSX has been growing much faster.  Customers are slow to adopt SDN in the data centre.
  4.  Meraki has been a huge success for Cisco, growing from $150MM to $500MM and is based on subscriber model. Robbins seems to have had a big hand in this.
  5. It could be that the board wants to focus on operations and sales during a time of change. Whitebox products in Ethernet, optical and routing could disrupt Cisco’s market position.

This article at Bloomberg has the following quote that bodes badly for employees suggesting that further headcount reductions and cost cutting will happen soon:

“Being the mathematician with a focus on numbers, we’re going to drive a level of operational rigor that maybe even is a little tougher than what you did, John,” the incoming CEO said.

Executive Departures

The Cisco executive ranks have been competing for the CEO position for several years and rumours abound of a dog fight to gain status and rank have been swirling around. Now that a candidate has been selected, most people expect that the exodus of long-term staff from Cisco will happen to senior executive ranks as they cash out or head to more interesting jobs in startups or competitors. This could create serious disruption internally as the balance of power between the internal fiefdoms/BUs shifts.

This competition has been both positive and negative for customers in terms of products and sales. Its possible or even likely that Cisco will go through serious disruption if large numbers of executives leave.

Something to watch for in the months ahead.

The EtherealMind View

This long overdue transition is a small start to adapting to a changing market. John Chambers isn’t leaving or going very far so its hard to see that much will change in the next year or so. The incoming CEO is strong on sales and operations which suggests that Cisco will continue to push hard to sell what is has today, continuing the fast follower strategy and avoiding innovation.

Ultimately, I believe that Cisco has limited control over the future as the data networking market goes through convulsive post-scarcity transition to a commoditized market. It faces unprecedented competition in nearly every market segment in which it operates such as Optical, SP edge, data centre, enterprise WAN and some markets are sunsetting such as Unified Comms, Videoconferencing and Cable Video.

Appointing a CEO who specialises in sales and process seems more like rent extraction from the existing business than a herald of real change at the company. And thats probably the best choice for shareholders in the short-term. Good luck and best wishes to the new CEO, Chuck Robbins.

Cue your own comments. Would be interested to hear what you think ?

References

Cisco Systems, Inc. – Cisco Board Names Chuck Robbins as Next CEO

Cisco taps veteran Robbins to replace Chambers as CEO | Network World

Is the Cisco Nexus 9000 is a Whitebrand strategy ?

By Filed Under: Blog, Cisco, Design

I was reviewing the non-ACI Nexus 9000 products this week and began considering that this is Cisco’s cisco-nx9000-whitebrandresponse to the disruptive impact of whitebox on switching sales.

Takeaways

  • Cisco Nexus 9000 now looks like whitebrand switch product following changes in the market.
  • The relatively low-cost of the product is attractive to Cisco’s loyal customers.
  • Market share by revenue is high but market share by ports or devices is changing.

Nexus 9000 (Non-ACI) Is Whitebrand/Britebox

There are three parts to a whitebrand or britebox Ethernet switch today.

  1. Whitebox Ethernet switch hardware
  2. branded operating system
  3. branded vendor support for software and hardware.

For most vendors, the whitebox Ethernet hardware comes from well known manufacturers based in China such as Quanta and Accton. The same companies who produce most networking hardware for Cisco, Juniper, Brocade etc.

Whitebox ethernet switch combinations

Nexus 9000 as Whitebox

The Nexus 9000 internal hardware is commodity merchant silicon. From what I can determine, the x86 processors, FPGA, memory, power supplies, fans etc are commodity components that are widely available on the open market. Broadcom StrataXGS/Trident switching silicon is used to rapidly bring the product to market in months instead of the usual five years for silicon build.

The ACI versions of the Nexus 9000 are not whitebrand product although they get benefits of commodity components, shared marketing and volume purchasing.


 whitebox-product-service-mix-20150421

 

(this graphic was updated at 20150421-1936 to correct some errors, my apologies to those impacted)

Operating System

The NX-OS is a Linux distribution. Cisco made a number of design choices to remove most or all Linux functions to make it appealing to Enterprise customers. Compared to Cumulus, BroadcomOS or PicoOS, NXOS represents a different design direction.

Hardware & Software Support

In my peer group there is consensus that Cisco’s Technical Assistance Centre (TAC) service & quality has been going downhill for some time but Cisco’s own customer satisfaction surveys remain high. There is no question that Cisco offers high quality but costly service capability on both hardware & software globally.

Pricing

The Nexus 9000 is cheaper than other Cisco switch products by a substantial amount and appears to replace a substantial amount of Nexus 7000/6000/5000 sales in the data centre market. I read this a sign of demand for a low-cost, simpler Cisco Ethernet switch product rather than feature improvement or acceptance of the ACI SDN platform.

Conflation or Compression

Some time back, I considered merchant silicon and whitebox products as separate issues. Over the last six months, my perception is changing:

  1. It has become clear that any Ethernet switch is mostly merchant silicon and commodity elements even if the switching chip is proprietary.
  2. My perception of the value of proprietary switch silicon is now low. Merchant silicon is coming that has flexible pipelines, dynamic protocol support and better performance from multiple suppliers.
  3. The operating system has become the focal point for features, functions and capabilities. That is, network features like MLAG, BGP/MPLS and orchestration features like Puppet, Chef & Ansible.

The outcomes of this perception shift are:

  1. Silicon will be critical in product selection and features in the same way that x86 CPU selection was important five years ago (but not today). It matters, but not that much. 
  2. operating system selection will be driven be features you are using in your network ie. STP, TRILL, MPLS, MLAG etc.
  3. operating system selection will be driven by the SDN platform in play i.e. Ansible, ACI, NSX, Nuage Networks, Big Switch all drive different operating system choices that subsequently drive hardware choices.

Cisco Whitebox

Cisco doesn’t need to embrace whitebox and its likely to be anathema the leadership team who have high levels of self belief in their “branding and value”

Cisco’s size means they manufacture more products than whitebox. But the market is shifting. The recent Delloro suggested that 7% of sales by revenue was whitebox in 2014. If you agree than Whitebox is one quarter the price, then this would suggest that whitebox volume would greater than 25% of ports or units shipped LAST YEAR.

Note:

The Cisco ACI SDN platform works best when using versions of the Nexus 9000 that has custom silicon to support extended features. These products would not fit the whitebrand model, are highly proprietary and restricted. They shouldn’t be confused with the discussion here.

IOS: show tcp vty

By Filed Under: Cisco, Operation

On Cisco IOS, this is a very useful command “show tcp vty xx” to show TCP statistics of the VTY session. If you think your terminal is running slow because of packet loss or delay then this command will provide visibility. The other cause is the CPU/Memory running slow if you don’t see any  errors on the TCP (as you can see below).

You will need to run the command for each VTY in use, but mostly I use for myself to check if its the network or the box.

It can also be useful for packet capture filtering of the console session since the source port of SSH Client is shown here.

If the link is lossy then look at the “datagrams” section and check the Smooth Round Trip Time (SRTT).

r1#sh tcp vty 0
tty66, virtual tty from host 192.168.10.130
Connection state is ESTAB, I/O status: 1, unread input bytes: 0
Local host: 10.216.4.10, Local port: 22
Foreign host: 192.168.10.130, Foreign port: 54596
Enqueued packets for retransmit: 1, input: 0 mis-ordered: 0 (0 bytes)
Event Timers (current time is 0x18C84473E):
Timer Starts Wakeups Next
Retrans 409 0 0x18C844865
TimeWait 0 0 0x0
AckHold 321 1 0x0
SendWnd 0 0 0x0
KeepAlive 824 0 0x18C853196
GiveUp 0 0 0x0
PmtuAger 0 0 0x0
DeadWait 0 0 0x0
iss: 2955761392 snduna: 2955805244 sndnxt: 2955805264 sndwnd: 65535
irs: 472603036 rcvnxt: 472609745 rcvwnd: 3908 delrcvwnd: 220
SRTT: 300 ms, RTTO: 303 ms, RTV: 3 ms, KRTT: 0 ms
minRTT: 4 ms, maxRTT: 401 ms, ACK hold: 200 ms
Flags: passive open, higher precedence, retransmission timeout
 keepalive running, nagle
TCB is waiting for TCP Process (139)
Datagrams (max data segment is 536 bytes):
Rcvd: 778 (out of order: 0), with data: 321, total data bytes: 6708
Sent: 503 (retransmit: 0, fastretransmit: 0), with data: 441, total data bytes: 43871
lontrm01#

 

 

 

If Cisco Could Be Split Up, What Could Go ?

By Filed Under: Blog, Cisco

Following the breakups of IBM and HP as they divest the low profit divisions and EMC under a some pressure to disband the Federation, the same question is often raised about Cisco but what could go ?

Cisco has generally pursued a good strategy of purchasing companies that are complimentary and  growth positive to its core business. While several acquisitions have failed (as is the norm) the survivors do integrate, support core revenue and leverage Cisco resellers. There are two product strategies that I see could be split from Cisco with minimal impact because of two fundamental shifts in technology.

IP Communications

Today, telephony and personal communications markets are dominated by smart phones and the dominant product, Apple iPhones, does not offer methods for integrating with Cisco IP Telephony platforms. Some employees no longer have desk phones or desks and rely on the mobile phone exclusively. This trend is progressively breaking down widespread use of legacy phone system. The cloud version, HCS, is well suited for ownership by a dedicated company that is free of the requirement to use Cisco hardware and bring the purchase pricing in line with competitors.

The rise of Internet-based services like WhatsApp, Skype and Facetime for personal use on smart phone is changing user behaviour. Its easier and safer to make personal call on Facetime than to use corporate resources. Over time, that this will reduce the perceived value of the dedicated endpoint or closed system represented by IP PBX systems.

For messaging, tools like Slack and Hipchat are forbears of new communication methods that use the network effect of the Internet to achieve deeper market penetration.

IP Video

The same logic applies to IP Video. It is unquestionable that Google Hangouts, GoToMeeting and WebEx are viable alternatives to traditional videoconferencing and this limiting the growth of the IP Video market. Sales will continue but growth seems unlikely.

The Cisco NDS business unit that provides in-network video services for carriers has limited opportunities for growth. Wider trends suggest that over-the-top Internet video from Youtube, NetFlix and other is more desirable than closed carrier video services such as offered by cable companies and limits the growth potential of the business.

Over The Top

A few years back these technologies were valuable for generating sales of network products (routing & switching) as part of the network architecture because bandwidth was limited and quality could only be assured in fully integrated private networks. What has changed is:

  1. Cheap and accessible bandwidth on the Internet has combined with updated video codecs to deliver ‘good enough’ video conferencing for low pricing or free. There is no longer a mandatory need to have deep network integration to deliver an acceptable experience.
  2. The consumer market is adopting smartphones that have voice and video solutions that are more convenient and reliable than Enterprise solution even though they operate at lower quality.

Divest Voice & Video

Because the growth of voice & video products seems limited because of fundamental trends in technology, I can envisage Cisco divesting with limited impact to core businesses. There is value in integration with other strategies, such as Cisco ACI or UCS although some sales would occur if the businesses remain integrated, the overall value is limited as the underlying growth of the respective markets is limited as consumer technologies take the growth.

Network Break Podcast

Network Break is round table podcast on news, views and industry events. Join Ethan, Drew and myself as we talk about what happened this week in networking. In the time it takes to have a coffee.

Packet Pushers Weekly

A podcast on Data Networking where we talk nerdy about technology, recent events, conduct interviews and more. We look at technology, the industry and our daily work lives every week.

Our motto: Too Much Networking Would Never Be Enough!

Find Me on Social Media