Editor’s note: The Skinny blog is written by WRAL TechWire editor and cofounder Rick Smith.
RESEARCH TRIANGLE PARK – It’s difficult to tune in a golf event or even some entertainment programs and NOT see ads from IBM touting its cloud computing efforts. But behind the video glitz there are problems that tech companies strive to avoid: Outages.
A rash of outages have hit IBM’s Cloud in recent months, raising some concerns about what’s happening as IBM prepares to split the company with its could focuse built around Raleigh-based Red Hat and the $34 billion acquisition of the Hatters two years ago to go all-in on cloud.
“With IBM betting the business on hybrid cloud, the constant issues with its own cloud can hardly be helping its sales team to convince customers that Big Blue has what it takes to help their businesses advance,” warns tech news site The Register.
No cloud is perfect, The Register notes. Yet …
“All clouds have outages,” the news site which follows IBM very closely says. “Google had a biggie last weekend when it reported “multiple Google Cloud Products experienced elevated latencies and/or errors due to an issue with Access Control Lists (ACLs) intermittently for a duration of 10 hours and 42 minutes.”
“But IBM has now had ten significant issues since April 3, by our count.”
That number certainly raised the eyebrows of one broadband expert who talked with WRAL TechWire on the basis that he not be identified.
“The recent IBM outages, especially their frequency, would be a cause for concern,” he said.
To be honest, he added: “IBM Cloud is relatively small compared with AWS [Amazon Web Services] and [Microsoft[ Azure.”
Figuring out how any cloud provider is doing is not easy, he adds.
“It’s difficult to assess the performance of the big public cloud providers. They don’t publish numbers for one thing. Another is that outages are usually not an all or nothing situation. Their services are so complex that portions may be unavailable for a time. It’s also possible for users to contract for specific types of services that don’t take advantage of redundancy built-in to the infrastructure and to take charge of their own reliability. AWS and others do quote reliability stats for their base services that are very impressive but impossible to verify.”
The Register has tracked IBM, though, noting that a “71” issue (“the rank it uses for incidents that see business-critical systems become unavailable” has occurred on:
- April 3
- April 20 (when two unrelated issues coincided)
- April 26
- May 20
- May 25
“Users may experience connectivity issues when trying to access the listed cloud services” IBM warned about a May 25 incident that struck Washington DC, Osaka, London, Dallas, Sydney, Tokyo, and Frankfurt.
The list of affected services is interesting since it lcudes IBM Watson – artificial intelligence which is the other big piece of the “new” IBM once the spinoff of its services group is complete.
Here’s the list from The Register:
- App ID
- Cloudant NoSQL DB
- Code Engine
- Continuous Delivery
- DNS Services
- Event Streams
- Hyper Protect Crypto Services
- Hyper Protect Virtual Server
- Hyper Protect DBaaS
- IBM Cloud Shell
- IBM Watson Machine Learning
- Mobile Foundation
- MQ in IBM Cloud
So is this a big problem for IBM?
“I have not heard any buzz in the research and education community. Again, this is probably a reflection of their size (or lack thereof),” the expert said.
“IBM touts their hybrid cloud platform that has components in IBM datacenters and on customer premises. It’s impossible to gauge the impact of their cloud incidents on customer operations. It’s probably zero for some customers and could be severe for others.”
One thing’s for sure: IBM sales reps are being quizzed about what’s happening.