Weekly internet health check, US and worldwide

ThousandEyes, which tracks internet and cloud traffic, provides Network World with weekly updates on the performance of three categories of service provider: ISP, cloud provider, UCaaS.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 Page 5
Page 5 of 14

On Oct. 7, Telia Carrier, a Tier 1 ISP headquartered in Stockholm, Sweden, experienced an outage affecting customers and downstream partners in the US, Spain, Germany, Austria, the UK, the Philippines, Japan, New Zealand, South Africa, the Czech Republic, Egypt, the Netherlands, India, and Canada. First observed around noon EDT, the outage initially centered on nodes in Ashburn, Virginia, and Sweden. Five minutes later the nodes in Sweden appeared to recover, while nodes in Newark, New Jersey, and London, England, began exhibiting outage conditions, increasing the number of countries and downstream partners affected. All except the London nodes appeared to recover 35 minutes after first being observed. The outage lasted an hour and 14 minutes and was cleared around 1:15 p.m. EDT.

Update Oct. 4

Global outages across all three categories last week decreased from 367 to 323, a 12% dip compared to the week prior. US outages remained the same at 161.

Globally, the number of ISP outages increased from 233 to 239, and in the US they increased from 115 to 132, up 15%.

Cloud-provider network outages worldwide dropped from 41 to 26, down 37%, while in the US they increased from four to six.

Both globally and in the US, collaboration-app network outages decreased by two, from eight to six.

There were two notable outages during the week.

On September 30, Oracle experienced an outage that affected customers and downstream partners interacting with Oracle Cloud services in the US. The outage was observed around 4:01 p.m. EDT and appeared to center on Oracle nodes in Frankfurt, Germany,  and Arlington and Ashburn, Virginia. Five minutes later, all nodes except those in Arlington appeared to recover. The outage lasted 7 minutes and was cleared around 4:10 p.m. EDT.

On September 29, NTT America experienced an outage that affected some customers and downstream partners across the US, Japan, and Hong Kong. It lasted around 18 minutes total, divided between two occurrences over a 25-minute period. The first was observed at 7:10 p.m. EDT and appeared centered on NTT nodes in New York, New York. It lasted four minutes and appeared to clear at around 7:15 p.m. EDT. Five minutes later, the second occurrence, lasting 14 minutes, affected the New York nodes plus nodes in Ashburn, Virginia. The outage to nodes in New York appeared to clear 10 minutes later, and the outage in Ashburn was cleared around 7:35 p.m. EDT.

Update Sept. 27

Global outages across all three categories last week increased from 276 to 367, a 33% increase compared to the week prior. In the US, outages increased from 116 to 161, a 39% increase.

Globally ISP outages increased from 186 to 233, a 25% increase, and in the US they increased from 87 to 115, up 32%.

Cloud-provider network outages jumped from 24 to 41, up 71%, while in the US they increased from three to four.

Collaboration-app network outages increased by eight worldwide and also by eight in the US.

Collaboration-app network outages reached the highest number they have reached this year. Both globally and in the US, they were up by eight, increasing from two to 10 worldwide, and from one to nine in the US.

There were two notable outages this week.

On September 21, Oracle experienced an outage affecting customers and downstream partners in the US, Germany, Finland, Singapore, Hong Kong, Australia, and Thailand. The outage was observed around 10:25 a.m. EDT and appeared to center on Oracle nodes in Frankfurt, Germany. Fifteen minutes later some of the nodes appeared to recover, reducing the affected countries to Finland, Hong Kong, and Australia. The outage lasted 24 minutes and was cleared at around 10:50 a.m. EDT.

On September 20, Comcast Communications experienced an outage that impacted downstream partners and customers in the US, Switzerland, China, and Hong Kong. The 18-minute outage was  observed around 4:50 p.m. EDT and appeared to center on Comcast nodes in Sunnyvale and Los Angeles, California; and Houston, Texas. Five minutes after it started, the disruption expanded to nodes in Santa Clara, California; New York, New York; Richmond, Virginia; Pittsburgh, Pennsylvania; and Denver, Colorado. Around 5:05 p.m. EDT, all nodes except those in New York appeared to recover The outage was cleared around 5:10 p.m. EDT.

Update Sept. 20

Global outages across all three categories during the past week increased from 209 to 276, a 32% increase compared to the week prior. In the US, outages increased from 85 to 116, a 36% increase.

Globally, the number of ISP outages increased from 149 to 186, up 25%, and in the US, they increased from 72 to 87, a 21% increase.

Cloud-provider network outages increased from 20 to 24, while in the US they increased from two to three.

Collaboration-app outages worldwide remained the same at two, and also the same in the US at one.

On September 14, Zayo Group, experienced an outage that affected some of its partners and customers in countries including the US, Hong Kong, Switzerland, and the UK. The outage lasted around 15 minutes, was first observed around 3 p.m. EDT, and appeared to center on Zayo Group nodes located in Philadelphia, Pennsylvania. Ten minutes later some of the node appeared to recover, reducing the number of impacted parties before the outage was cleared around 3:20 p.m. EDT.

On September 13, Microsoft experienced a network outage that impacted downstream partners and access to services running on Microsoft environments. The 29-minute outage was first observed around 1:25 a.m. EDT and appeared to be centered on Microsoft nodes in Des Moines, Iowa. Five minutes later, nodes located in Chicago, Illinois, showed outage conditions, resulting in an increase in the number of impacted partners. Around 1:35 a.m. EDT, nodes in Cleveland, Ohio, also showed outage conditions. Twenty five minutes after the outage was first observed, the nodes in Des Moines and Cleveland appeared to recover about 25 minutes into the outage leaving just the Chicago nodes as the only ones still out.The outage was cleared around 1:55 a.m. EDT. Given the duration and timing relative to the location of the nodes at the center of the outage, it is likely to have been a maintenance exercise.

Update Sept. 13

Global outages across all three categories last week increased by one, from 210 to 209. In the US, they decreased from 90 to 85.

ISP outages grew worldwide from 140 to 149 and grew in the US from 68 to 72.

Globally, the number of cloud-provider network outages stayed the same at  20, while in the US they dropped from seven to two.

Collaboration-app network outages remained the same at two, and in the US the number rose from zero to one.

There were two notable outages.

At 1:45 p.m. EDT on Sept. 7, Amazon’s network experienced a disruption affecting downstream partners and customers in the US, Japan, South Korea, Ireland, and the Philippines. It lasted around 18 minutes distributed across two occurrences over a 35-minute period. The first instance appeared to center on Amazon nodes in Incheon, South Korea, and lasted around nine minutes affecting users in the US, the Philippines, and South Korea. Around 2:10 a.m. EDT, approximately 15 minutes after the first occurrence cleared, a second occurrence was observed, also lasted around 9 minutes, and appeared to again center on Amazon nodes located in Incheon, South Korea. It  affected service in South Korea, Japan, and Ireland. The outage was cleared around 2:20 a.m. EDT.

On Sept. 6, Cogent Communications, experienced a series of outages over a half-hour period that affected downstream providers and customers in multiple countries, including the US, Spain, Greece, Germany, Luxembourg, the Ukraine, and Portugal. It lasted 13 minutes divided between two occurrences over a 25 minute period. The first occurrence was observed around 7:15 p.m. EDT centered on Cogent nodes in Sacramento, California, and Bilbao, Spain, and lasted around  nine minutes. The Cogent environment was stable for 15 minutes then experienced a series of four-minute outages observed on Cogent nodes in Bilbao, Spain. The outage was cleared around 7:45 p.m. EDT.

Updated Sept. 6

Global outages across all three categories last week decreased from 286 to 210, down 26%, with most of the decline coming from improved performance in the US where the total outages dropped from 145 to 90, down 37%.

ISP outages globally dropped from 214 to 140, a 35% decrease. In the US they decreased from 119 to 68, a 43% drop.

Globally, cloud-provider outages increased from 17 to 20, while in the US they increased from five to seven.

Worldwide, collaboration-app network outages dropped by from four to two, with the decrease coming from improved performance in the US where outages dropped from two to zero.

There were two notable outages during the week.

At 2:15 a.m. EDT on Sept. 1, Microsoft experienced a 29-minute outage affecting downstream partners and services running in Microsoft environments that appeared centered on nodes in Des Moines, Iowa; Chicago, Illinois; and Spokane, Washington. Five minutes in, nodes in New York, New York; Los Angeles, California; Amsterdam, the Netherlands; Sydney, Australia; Portland, Oregon; and Cleveland, Ohio; also began exhibiting outages. Around 2:25 AM EDT, the nodes in New York, Sydney, Portland, Chicago, Spokane, and Cleveland appeared to recover. The outage was cleared around 2:45 AM EDT. Given the duration, timing, and the location of the nodes, it is likely to have been a maintenance exercise.

At 12:05 a.m. EDT on Sept. 2, Cogent Communications experienced an outage affecting multiple downstream providers and customers in the US, Brazil, Australia, Canada, Singapore, South Africa, and the K. It lasted around 14 minutes and appeared to be centered on Cogent nodes in New York, New York. Five minutes into it some of the nodes appeared to recover, reducing the number of countries affected to just the US. The outage was cleared around 4:20 AM EDT.

Updated Aug. 30

Global outages across all three categories remained fairly steady last week, increasing from 284 to 286. In the US, outages bumped up from 99 to 145 outages, a 46% increase.

Globally, the number of ISP outages increased from 193 to 214, up 11%, while in the US, they increased from 68 to 119, a 75% jump.

Worldwide cloud provider outages dropped from 33 to 17 outages, down 48%. In the US they also dropped, from 19 to 5, a 74% decrease.

Collaboration-app network outages worldwide remained at four, and decreased from three to two in the US.

There were two notable outages during the week.

On August 24, Cogent Communications experienced an outage that impacted both downstream providers as well as customers in countries including the US, New Zealand, Ukraine, Spain, Mexico, Luxembourg, Hong Kong, Singapore, the UK, Republic of Korea, Portugal, Japan, Germany, Greece, China, Philippines, Brazil, Australia, France, and Argentina. First observed around 5:35 a.m. EDT, it appeared initially centered on Cogent nodes in Los Angeles, California. Five minutes later, the location of nodes terminating traffic increased to include Phoenix, Arizona; Las Vegas, Nevada; San Francisco, San Jose, and Fullerton, California; El Paso and Houston, Texas; Atlanta, Georgia; and Hong Kong. The outage lasted 59 minutes and appeared to coincide with Cogent planned-maintenance that involved code upgrades on their infrastructure in the Los Angeles area. The outage was cleared around 6:35 a.m. EDT.

Also on August 24, Zayo Group experienced an outage affecting some of its partners and customers in countries that included the US, Canada, Singapore, India, and Hong Kong. The outage was first observed around 3 a.m. EDT and lasted around 18 minutes. Initially it appeared centered on Zayo Group nodes in Seattle, Washington, and Los Angeles, California. Five minutes later the location of nodes exhibiting outages expanded to include San Diego, California, and Phoenix, Arizona. For the next 10 minutes, the number of nodes terminating traffic began to clear until around 3:15 a.m. when the only outage conditions were in Phoenix. That outage cleared around 3:20 a.m. EDT.

 

Updated Aug. 23

Worldwide outages across all three categories last week increased from 275 to 284 compared to the previous week. In the US, outages increased from 85 to 99.

Globally, the number of ISP outages decreased from 200 to 193, but in the US they increased 11% from 61 to 68.

Cloud provider network outages worldwide from 29 to 33 outages, and in the US, jumped from three to 19.

Globally and in the US, Collaboration-app network outages decreased globally from five to four and in the US from four to three.

There were two notable outages during the week.

At 4 a.m. EDT on August 18, Rackspace Technology experienced an outage affecting some of its customers and in countries including the US, Switzerland, Canada, and Germany. The outage lasted around 30 minutes in total, divided among six occurrences distributed over a period of 75 minutes. The first appeared to center on Rackspace nodes in Washington, DC, and lasted 14 minutes before appearing to clear at around 4:15 a.m. EDT. Five minutes later a series of four outages were observed, each lasting around four minutes with a period of five minutes between each, still appearing to center on nodes in Washington. The final occurrence was observed around 5:10 a.m. EDT, before clearing around 5:15 a.m. EDT. 

At 6 p.m. on August 18, NTT America experienced an outage affecting customers and downstream partners across the US and Canada. It lasted about six minutes and was divided into two occurrences over a 15 minute period, appearing to center on NTT nodes in Seattle, Washington. It was cleared around 6:15 p.m. EDT.

Updated Aug. 16

The total number of outages worldwide across all three categories increased from 201 to 275 during the past week, up 37%, while in the US they rose from 65 to 85, up 31%.

Globally the number of ISP outages increased from 149 to 200, a 34%, rise and in the US went up from 51 to 61, a 20% increase.

Worldwide cloud-provider network outages jumped from 11 to 29, a 163% increase, and in the U.S., increased from two to three.

Globally, collaboration-app network outages increased from one to five outages, while the number in the us went from zero to four.

Related:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 Page 5
Page 5 of 14