Barclays snubs public cloud giants and hardware rivals for HPE GreenLake private cloud Hewlett Packard Enterprise's direct salesforce has inked a 10-year contract with Barclays Bank to provide a global private cloud via its GreenLake platform. Under the agreement, understood to be worth many millions of dollars, HPE will refresh the infrastructure in multiple Barclays' data centres and co-location facilities …
Barclays backed off AWS because they spent a rumoured-to-be-nine-figure sum to AWS in the space of a couple of years attempting a "Cloud Native Services For Everything" strategy and found themselves having spent all that money but having migrated the sum total of fuck all in that time.
Plenty of new/organic services going on up there - especially in the retail banking space where customer proximity and resilience really matter - but the grand strategic plan to move on premises stuff to cloud using excusively AWS native services failed and failed badly - the Barclays "Cloud Architects" drunk the AWS kool aid hard and regret it. Mostly from recently-acquired new jobs elsewhere.
This deal isn't as all-encompassing as HPE would make out. Barclays have always been a significant HPE customer - most of their kit is HPE kit and they buy plenty of software from them - so this is mostly a restructuring of existing deals onto a new PAYG pricing structure. There's no appetite to replatform shedloads of stuff onto HPE's half-baked MapR castoffs, and the reality is Barclays already have huge on premises "private clouds" from VMWare and Openshift, which HPE aren't going to be challenging any time soon.
Given your massively accurate portrayal of Barclays cloud journey (2 years?, suggest you check re:Invent 2017) I will make sure to take the rest of your well reasoned arguments regarding what has been achieved and how they intend to go forward with muti-cloud with the same level of authority
Well what was actually said was that they spent a rumoured-to-be-nine-figure-sum in a couple of years. If we're splitting hairs on timelines the failed cloud strategy itself was launched in 2016, which makes the general lack of return on investment all the more damning. Which is a big part of why they flew that aussie lot in to run the tech & data show.
I did some consulting with Barclays a few years ago. Not only did they have no control over business applications they had no control over projects and programmes either. They couldn’t even figure out what their IT spend was, let alone number of projects, business cases, etc.
I give it two years max.
"Delivering an enhanced personalized banking experience" for customers."
Based on a number of issues and cases I've read about recently it seems they could have more efficiently improved this by employing and adequately training more front line service staff. In this instance, and similar to other large players "enhanced personalized experience" most likely means more useless chat-bots, web based self-help which never gives you any actual help, and more automated phone routing which will invariably cuts you off just when you think you are getting somewhere.
All of the above. It's as much a brand as anything else - at a technical level it's really an orchestration/monitoring/billing engine for a bunch of other cloud-ish techs. So you can do a bit of SDN, a bit of container orchestration, maybe some job management - it even talks to AWS or Azure.
The question is whether it's going to do any of that well. A lot of the tech actually being orchestrated was bought up at bargain basement prices from failed firms (e.g. MapR + Bluedata underpin Ezmeral). So there's two big risks. One is whether the various bits work and whether they work well together.
The other is whether the various bits play nicely with other peoples' products. It's all well and good shipping a Kubernetes orchestrator but if it has a different API to native K8s and a different API to AKS/EKS/OCP etc you'll have a hard time finding anything else to run on it.
Which a more cynical person would assume to be the point.
That’s actually wrong, we have Greenlake for IaaS. HPE manages servers, storage, network and VMware vSphere. We manage VM provisioning and OS. We pay for a baseline 80% of the capacity and have on demand access to a 20% capacity buffer. We can scale up and down monthly. The net cost is cheaper than AWS, but more expensive than a pure DIY play. But if you’re having trouble staffing it’s just moving the cost from employees to Greenlake services.
It wouldnt surprise me if there is a SEPARATE banking systems just to "process" the very special accounts away from prying eyes and open S3 buckets that might "accidentally" leak into the hands of journos. And we get the Barclays papers leak anytime soon if they went with any of the Public clouds.
Gotta keep those dolla dolla bills laced with polonium flowin'
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
