Tag Archive for Networking

Should I Care About 768k Day?

Why Should I Care About 768k Day?If you are of a certain age, you remember Y2K. While I was not rewriting COBOL programs, I played my part. I spent the last half of 1999 scheduling after-hours downtime to update Compaq 1900 and 2500 servers with BIOS updates on a floppy disk. Hoping and praying the servers would come back up after the floppy disk stopped grinding. As I recall only two Compaq Proliant 2500‘s failed the BIOS upgrade and only one was DOA.

All the fun of Y2K was because memory space was too small to accommodate the fancy new year 2000 without thinking it was 1900. Now a similar memory size problem could cause internet disruptions very soon. The problem is called 768k Day.

768k Day is when the size of the global BGP routing table is expected to exceed 768,000 entries. Anthony Spadafora at TechRadar explains that on August 12, 2014, a similar problem, occurred after Verizon (VZ) advertised 15,000 new BGP routes to the internet. Verizon’s actions caused the global BGP routing table, a file that holds the IPv4 addresses of all known internet-connected networks, to exceed 512,000 causing the 512K Day crisis.

Over flowingThe TechRadar article explains that in 2014, ISPs and others had configured the size of the memory for their router TCAMs (ternary content-addressable memory) for a limit of 512K route entries and some older routers suffered memory overflows which led their CPUs to crash. These crashes created significant packet loss and traffic outages across the internet with even large provider networks being affected. ZDNet says companies like Microsoft, eBay, BT, Comcast, AT&T, Sprint, and Verizon, were all impacted by 512K day

Engineers and network administrators rushed to apply emergency firmware patches to set a new upper limit which in many cases was 768k entries. The seeds of the 2019 768k  crisis were sown.

preventative maintenanceMr. Spadafora speculates that in 2019 most of the large providers who felt 513K day’s impact have likely updated and maintained their infrastructures reasonably well which could lead to fewer outages. He says that there are still ‘soft spots’ smaller ISPs, data centers, and other providers who are part of the Internet’s fabric where maintenance on legacy routers and network equipment can be neglected or missed more easily.

These are the places that most likely see some issues or outages due to 768k Day. These outages will create significant packet loss and traffic outages that could have a ripple effect and sweep upstream and affect larger provider networks. Alex Henthorn-Iwane at network intelligence firm ThousandEyes writes,Given the sheer size and unregulated nature of the Internet, it’s fair to say that things will be missed.

rb-

To prepare for any potential disruptions, it is a good idea to perform some preventative maintenance on any routers that receive full internet routes. Jim Troutman, Director at the Northern New England Neutral Internet Exchange (NNENIX) told ZDNet,

The 768k IPv4 route limit is only a problem if you are taking ALL routes. If you discard or don’t accept /24 routes, that eliminates half the total BGP table size.

There is still a little time left before 768K day, at 2019-06-21 16:00 UTC 06/21/2019 the Regional Internet Registry for Europe, the Middle East, and parts of Central Asia (RIPE) reports that 86.9% of the IPv4 BGP tables they monitor are below 768K. Click here for current results

What is the big deal? Network intelligence firm ThousandEyes points out that there are many outage events that happen every day, especially on the fringes of the Internet. The number of garden variety outages could get amplified because of 768k day-related issues over the next few weeks.

Aaron A. Glenn, a networking engineer with AAGICo Berlin told ZDNet,

Cisco 6509The Cisco 6500/7600 product line was extremely popular for an exceptionally long time in many, many places,” so don’t be surprised if some networks go offline because they forgot about 768k Day and didn’t prepare.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.

Why Shielded Cables?

Why Shielded Cables?With high bandwidth applications emerging in and out of the data center, the need for shielded cable has increased. Many of these new installations are in harsh environments.  Robotic equipment, motors, generators, air conditioners, fluorescent lights, printers, and they often generate a large amount of EMI (Electromagnetic Interference) and RFI (Radio Frequency Interference). You may need to install shielded cables in these harsh environments.

interference will cause increased errorsEMI and RFI can cause crosstalk between circuits and interfere with data transmission on a copper cable. The interference will cause increased errors resulting in mire network traffic due to packet retransmissions, and downtime.

Shielded cables reduce interfernece

EMI is an unwanted signal that is induced into the cable. EMI typically comes from a source that is external to the cable, such as an electrical cable or device. Cables can be both a source and receiver of EMI. As a source, the cable can either conduct noise to other equipment or act as an antenna radiating noise. As a receiver, the cable can pick up EMI radiated from other sources.

RFI is a disturbance that affects an electrical circuit due to either electromagnetic conduction or radiation emitted from an external source. Conducted RFI is unwanted high frequencies that ride on the AC waveform. Radiated RFI is emitted through the air.

Ethernet cables can be shielded to deal with EMI and RFI.  The shield surrounds the inner signal carrying conductors. Shielded Ethernet cables can deal with interference in two ways. It can reflect the energy, or it can pick up the interference and conduct it to ground. Both methods use shielding to cut the EMI and RFI reaching the twisted pairs located under the shielding. Whether the EMI/RFI is reflected off the shield or “rides” the shield to ground, some energy can still pass through the shielding, but since it is so highly attenuated it will not cause interference.

Types of shielded cables

Two basic types of shielded cables are available. The first are cables with an overall shield known generically as STP. STP cable may be made either with a foil or a braid for the shield. STP cables with an overall foil are often known as FTP. In practice, FTP and STP cables may be interchanged with no apparent difference in performance according to BlackBox

Cable types
The second type of shielded cable is S/STP. S/STP cables have a shield around each individual pair and an overall outer shield. BlackBox says the purpose of the inner shields is to cut the Alien Cross Talk parameter in CAT7 and CAT6a systems.On both STP and S/STP the primary task of the outer shield is to resist external RF noise such as electrical spikes.  S/STP cables may also have an extra braid for strength and to simplify connection to the metal shields around the connectors.
Cable constructionTo reduce EMI/RFI interference, shielded Ethernet cables must use shielded connectors to maintain the benefits of STP cabling. High-quality shielded cable includes a drain wire to provide grounding that cancels the effects of EMI and can ease termination of the cable shield for crimping or soldering.

Redi WattProper Ground connections

The shielded cable system must have proper ground connections for the shields. Incorrect grounding opens the possibility for ground loop currents and associated interference to the Ethernet signal. In the worst cases with no proper grounding, the shields can actually act as antenna broadcasting high-frequency signals out into the environment interfering with electronic equipment and allowing external detection of the Ethernet data.

Standards bodies have requirements for shielded Ethernet systems. The American National Standards Institute (ANSI) specifics how shielded communications cables are properly grounded. Normal practice says the cable shields should be grounded in the telecommunications closet (TC). Typically, the cable shield is grounded through the connector to the patch panel. Then the panel is grounded to the rack, which is grounded to the telecommunications grounding busbar in the TC. Do not ground the cable shield at the work-area outlet. Fortunately, many of today’s shielded outlets automatically connect to the patch panel’s ground so there’s no need to set up ground paths for each cable.

ISO/IEC requires that shielding must be complete for an entire channel, shielded and shielded cables, connectors or network controllers should not be mixed.

Cable Types

Category cable types comparison.
Max Data RateMax TX RateMax LengthSheildingYear IntroducedStatus
Cat 310Mbps16MHz100mNo1983Obsolete
Cat 5 1,000Mbps100MHz100mOptional1995Obsolete
Cat 5e1,000Mbps250MHz100mOptional2001Obsolete
Cat 6 10,000Mbps500MHz100mOptional2002
Cat 6a10,000Mbps500MHz100mOptional2008
Cat 7 10,000Mbps600MHz100mRequired2002Not recognized by TIA
Cat 7a10,000Mbps1GHz100mRequired2010Not recognized by TIA
Cat 8.140,000Mbps2GHz30mRequired2016Backward compatible with Cat 6A
Cat 8.240,000Mbps2GHz30mRequired2016Not recognized by TIA

rb-

interference is not a new problem. It dates back to the earliest communication systems. In 1881 Alexander Graham described the interaction between many twisted pairs in US Patent 244,426.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.

IPv4 Update

IPv4 UpdateThe IPocalypse struck the United States in 2015 and three years later – nobody cares. The end of IPv4 was going to mean the Internet Protocol version 6 (IPv6) takes over the world. Well, recent updates say IPv4 is still the dominant protocol on the Internet.

IPv4 number trading between private partiesIPv4 number trading between private parties has proved to be an effective means of extending the life of IPv4 by redistributing previously allocated IPv4 numbers. Trading between private parties is very active in North America, Europe, and the Asia-Pacific regions. This has allowed IPv4 network operators to support and extend their IP networks with excess unused supply through the IPv4 market.

Janine Goodman, Vice President and co-founder of Avenue4 LLC., a Washington DC-based IPv4 broker and advisory firm posted a 2018 Q3 update on the IPv4 market for CircleID. In the update, the author noted that during Q3 of 2018 there is still a voracious appetite for IPv4 numbers – 18 million IPv4 numbers were transferred in the quarter. There were nearly 42 million IPv4 addresses changing hands in the 2018 year to date, a 160% jump compared to 2017

will not hit the 50% mark until sometime in 2020The article states that IPv6 adoption in the U.S. (based on Google user stats) fell after the beginning of the year and has yet to recover. Ms.Goodman cites predictive models which suggest that U.S. IPv6 adoption will not hit the 50% mark until sometime in 2020. Globally, IPv6 adoption has been slow, peaking at 21.5% during weekdays and 25% during weekends. The data from Avenue4 confirms that IPv4 continues to be the dominant Internet protocol.

For those firms purchasing IPv4 addresses, the most common IP block size is the /24 (256 addresses), followed by the /16 block (65,536 numbers). Nearly 90% of those /16 blocks were transferred to large block buyers. The /17 (32,768 addresses) and /18 (16,384 addresses) are also popular as large block buyers are increasingly willing to accept a collection of smaller non-contiguous ranges from sellers.

prices are being driven up by fierce competitionBlock prices will rise over the next 6-12 months. The article reports that most block sizes are north of $17.00 / number with larger blocks reaching and occasionally exceeding $20.00 / number. Avenue4 says fierce competition among large block buyers is driving unit prices up. Until this demand is met, pricing should continue to escalate. IPv4 prices were $11.25 / address when I first wrote about Microsoft’s purchase of Nortel’s IPv4 addresses in 2011.

rb-

The switch to IPv6 is being slowed by two factors. The first is network address translation (NAT) which has become better understood and implemented. The other is the evolution of the buying and selling of IPv4 addresses, led by firms like Avenue4 and IPv4 Brokers.

One of the knocks against moving to IPv6 is CAPEX and OPEX costs. But neither really holds water anymore. The data from Avenue4 says that firms are willing to pay over $1.3 million for a \16 block. Firms could leverage $1.3 million to update to IPv6.  IPv6 is fully built into modern operating systems and networking hardware. Buy the right devices during your regular update cycles.

Microsoft (MSFT) recently decided to embark on the tricky transition from IPv6 and IPv4 or ‘dual stack’ to IPv6-only, which Microsoft believes will solve its problems with IPv4 shortages and enable simpler network management.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.

Switch Sales Stalled

Switch Sales StalledThe stats for sales of network switches are in for Q4 2017. Only one of the top 5 networking vendors was able to squeeze out a small gain in switch sales. The data comes from New York-based NPD. NPD tracks monthly network switch sales data from the sales channel, distributors, and resellers in North America.

The article on CRN notes that the total number of switches sold through the channel in the quarter was 514,095. The number is up slightly from 510,822 in the fourth quarter of 2016, according to NPD. Here are the five vendors that sold the most switches through the channel in the fourth quarter, according to NPD.

D-Link Systems

D-Link logoTaiwan-based D-Link Systems (2332:TT) sold 25,259 switches during the fourth quarter, according to NPD statistics. That total kept the company steady with the same period in 2016 when it sold 25,277. D-Link did not have a switch model among the top 10-selling units during the quarter. Its market share was unchanged at 4.9%, CRN said.

TP-Link switch sales

According to NPD’s data, of all the five best-selling switch brands, TP-Link saw the steepest decline during this period. The company based in Shenzhen, China sold 26,023 switches in Q4 ’17 compared with 29,798 in Q4 ’16. That’s a 12.7 percent year-over-year decrease. There is one bright spot for the firm, the article reports that the company’s TLSF1005D Ethernet switch was the third-best-selling unit during the quarter. But that was not enough to prevent a market share decline from 5.8 percent in 2016 to 5.1 percent in 2017.

Hewlett Packard Enterprise switch sales

HPE LogoThe news from NDP is not good for former networking giant Hewlett Packard Enterprise (HPE) either. The Palo Alto, CA-based firm saw a 1.8 percent decline in switches sold from 55,923 in Q4 ’16 to 54,941 switches in Q4 ’17. The quarter’s total was enough for a 10.7 percent market share, down slightly from the year-ago period. No HPE switch models were among the top 10 for the quarter, according to NPD.

Netgear sales

CRN reports that sales also slipped for Netgear. The number 2 switch company saw its market share dip from 18.3% to 17.9% year over year. The California-based firm sold 92,274 switches through the channel in the fourth quarter, down slightly from the 93,531 it sold in the same period a year ago, NPD said. Netgear had four switches in the top 10-best-selling switches during the quarter, including the top two models, the FS105 and GS105NA five-port models.

Cisco switch sales

Cisco (CSCO) was able to hold on to the #1 switch vendor position according to NDP. It sold 225,051 units during the period, a 5.7 percent increase that boosted the company’s market share to 43.8 percent from 41.7 a year earlier. Six of the top 10 best-selling switches in the quarter were Cisco Catalyst‘s led by the WS-C2960X 24– and 48-port models.

rb-

What happened to the network switch market? It’s still reeling from the 2007/08 recession and the Wi-Fi takeoff. Other than the Cisco switches, most of the top switch models sold were unmanaged, desktop switches limited to 100 Mbps uplinks. These types of switches make it OK to randomly add an unauthorized switch at the desktop and POOF there does your data. These desktop switches with their limited feature set don’t include Spanning Tree, so users can create a network loop and take down the whole network segment.

Not much to shout about.

Where are the vendors? Brocade? Extreme? Juniper? Dell? I am old enough to remember when switch manufacturers had a #2 strategy. 3Com, Lucent, Bay/Nortel all came into my office and said they wanted to #2 – now they are gone.

Related article

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.

Whats Up With Cisco?

Whats Up With Cisco?What is up with Cisco? Their fiscal results for 2017 Q3 showed revenue of $11.9 billion, a 1% decline in revenue, compared to last year. This is the 6th consecutive down quarter. The networking goliath also issued downward guidance for 2017 Q4. They estimated a revenue declines of 4-6% year-over-year.

Cisco logoOn the earnings call, Cisco CEO Chuck Robbins blamed several factors for the lower guidance. He cited:

  • “A pretty significant stall right now” in the U.S. federal public sector.
  • Service provider revenues were down in Mexico.
  • United Kingdom business is being dampened by currency issues.
  • Middle East, there is “pressure… relative to oil prices.”

Cisco layoffs

Then there are the layoffs. Cisco buried the announcement in a footnote in the company’s SEC 8-K report that 1,100 more layoffs are coming. That is on top of the 5,500 announced in August 2016.

In May 2017, we extended the restructuring plan to include an additional 1,100 employees with $150 million of estimated additional pretax charges.

Cisco layoffs

According to SDXCentral, the Cisco CEO stressed several times on the earnings call, that the company is transitioning to more software and subscription-based business. He declared,

I am pleased with the progress we are making on the multi-year transformation of our business.

These weak financial results and the move to a subscription-based business have fed speculation about the future Cisco business model. TechTarget speculates that Cisco may go so far as to separate the Network Operating System (NOS) from the hardware. They contend the move would be a dramatic departure from Cisco’s traditional business model of bundling high-margin hardware with its NOS. The author believes that market trends will likely force the vendor to release an open NOS.

Open NOS

Cisco 3750 switchTechTarget cites reports from The Information that a hardware-independent NOS called Lindt is coming. Reportedly Lindt will run on a white box powered by merchant silicon. According to the article, a number of market trends are driving the move to a hardware-independent NOS.

The first market trend forcing Cisco’s hand is the company’s declining dominance of the Ethernet switch market. Since 2011, the company’s share has dropped from 75% to less than 60% last year, according to the financial research site Trefis. The decline is important to Cisco’s bottom line. Switches accounted for 40% of Cisco’s product sales in 2016, 30% of net revenues, and 20% of the company’s $162 billion valuation.

Infrastructure as a ServiceCisco’s weakening performance in switching is tied to the second market trend forcing Cisco to release a hardware-independent NOS. Its customers are turning to public cloud providers, Amazon (AMZN) Web Services, Microsoft (MSFT) Azure, and IBM (IBM) SoftLayer, for their IT infrastructure. The more enterprises subscribe to infrastructure as a service, the less networking gear they need in their data centers.

Cloud computing

The shift to cloud providers is found in the latest numbers from Synergy Research Group. Revenue from public cloud infrastructure services is growing at almost 50% a year. In the fourth quarter of last year, revenues topped $7 billion.

 cloud providers are building open networking hardware and softwareThe third market trend forcing Cisco to a hardware-independent NOS is enterprises that were Cisco’s largest customers are now competitors. Enterprises and cloud providers are building open networking hardware and software to replace inflexible proprietary systems that lock them in. Those companies include large financial institutions, like Bank of America, Goldman Sachs, and Fidelity Investments. As well as communication service providers, AT&T (T), Deutsche Telekom, and Verizon (VZ).

The technology shift is driving an enormous amount of spending on IT infrastructure. Worldwide spending on public and private cloud environments will increase 15% this year from 2016 to $42 billion, according to IDC. Meanwhile, spending in Cisco’s core market of traditional infrastructure for non-cloud data centers will fall by 5%.

White boxes

Arista NetworksWhile Cisco is ignoring the trend away from proprietary hardware, the article says Cisco’s rivals are embracing it. Juniper (JNPR) and Arista (ANET) have released versions of their NOS for white boxes favored by cloud providers and large enterprises. Both companies reported year-to-year revenue growth in switching last year. Even Cisco’s patent lawsuit against upstart Arista was set back by the courts.

Rohit Mehra, an analyst at IDC hypothesized that Cisco’s resistance to change is likely due to fear that giving customers other hardware options would accelerate declining sales in switching. “There would be potentially some risk of cannibalization in the enterprise space,” he added.

Cisco insists its customers are not interested in buying networking software that’s separate from the underlying switch. The Cisco spokesperson told TechTarget:

TCisco insists its customers are not interestedhe vast majority of our customers see tremendous value in the power and efficiency of Cisco’s integrated network platforms, and the tight integration of hardware and software will continue to be the basis of the networking solutions we offer our customers

TechTarget adds that Cisco doesn’t say the article is wrong. Instead, the company falls back on a corporate cliché for refusing to discuss a media report. “We don’t comment on rumor or speculation,” a Cisco spokesperson said.

The networking market is evolving away from the hardware that Cisco depends on for much of its valuation. Cisco will resist changing its market approach for as long as possible. But in the end, the company will have to become a part of the trend with an open NOS capable of running on whatever hardware the customer chooses.

Cisco’s own problems

Rather than change its model for selling networking gear, Cisco has spent billions of dollars on acquisitions over the last few years to create software and subscription-based businesses in security and analytics. But Cisco’s software push has yet to pay off with 5 conservative down quarters.

Finally, Cisco just recently patched a flaw in IOS software that affected more than 300 models of its switches. Despite issuing an advisory on March 17, Cisco did not release the patch for this vulnerability until May 8, 2017. The Cisco vulnerability was part of the Vault 7 WikiLeaks dump of alleged CIA hacking tools.

Alleged CIA hacking toolsThe vulnerability, rated a critical 9.8 out of 10 by the Common Vulnerability Scoring System, is in the Cluster Management Protocol, or CMP. could allow a remote, unauthenticated attacker to reload devices or execute code with elevated privileges. This vulnerability can be exploited during Telnet session negotiation over either IPv4 or IPv6.

Related articles

 

Ralph Bach has been in IT long enough to know better and has blogged from his Bach Seat about IT, careers, and anything else that catches his attention since 2005. You can follow him on LinkedInFacebook, and Twitter. Email the Bach Seat here.