Datacenter – Dataconomy

Mesosphere Grabs Funding to Deploy World’s First Datacenter Operating System

Dan Gray — Tue, 09 Dec 2014 08:59:33 +0000

Mesosphere, the San Francisco based start-up, has rolled out a one of a kind Data Center Operating System (DCOS) based on the open source Apache Mesos project and to help it with further development, the outfit has garnered $36m in funding.

The announcement came earlier this week as Khosla Ventures led a fresh round, along with Andreessen Horowitz, Fuel Capital, and SV Angel. Raking in $10 million earlier this year the total funds raised is now $48.85 million.

Brad Silverberg, co-founder of Fuel Capital explained, “Virtual machines are a step along the way, but the new breed of datacenter services run across fleets of machines. The Mesosphere DCOS is the first operating system to unleash the full power of the cloud — you can run existing applications, but you can also build new applications, bringing automation to processes that otherwise require too much human oversight and simply do not scale.”

The Mesosphere DCOS “pools compute resources and automates common operations,” to providing developers and operators with “an interface to automate allocation and deallocation of datacenter resources to run today’s most popular distributed applications and services.” Twitter, Airbnb and Hubspot are some of the sites running on Mesosphere tech.

According to their press release, key features of DCOS include:

A distributed systems kernel with enterprise-grade security, based on Apache Mesos.
A set of core system services, including a distributed init system (Marathon), distributed cron (Chronos), service discovery (DNS), storage (HDFS), and others, all capable of launching containers at scale.
A user interface, including a command line for controlling the Mesosphere DCOS and visual tools for understanding how your datacenter is operating.
Killer apps and an ecosystem of datacenter services, including Apache Spark, Apache Cassandra, Apache Kafka, Apache Hadoop, Apache YARN, Apache HDFS, and Google’s Kubernetes, supported by Mesosphere and third parties as native services on the Mesosphere DCOS.
A public and private repository, which allows operators to deliver third-party and custom-built datacenter services to their business with a one-command install.
An API and software development kit (SDK) that lets programmers develop against datacenter like it’s one big computer, which aids quickly building applications and services that scale with built-in high availability and fault tolerance.
Support for a wide variety of platforms, including:
- All modern versions of Linux: Redhat, CentOS, Ubuntu, and CoreOS
- On-cloud: Amazon, Google, DigitalOcean, Microsoft, Rackspace, VMware
- On-premise: Bare metal, VMware, OpenStack

“This is an operating system being built by distributed systems developers, to tackle the biggest inefficiencies that we see in how today’s applications and services interface with compute resources,” points out Florian Leibert, CEO and Co-Founder at Mesosphere.

“From resource sharing, to handling failure, and everything in between – the mission of the DCOS is to bring automation where automation makes sense, and to give developers and operators much finer grained control of their environments, while freeing them from low level plumbing,” he further added.

It has a workforce of nearly 40 people and is set to rise to 120 by the end of 2015, reports VentureBeat.

Follow @DataconomyMedia

(Image credit: John McStravick)

MIT Researchers Find Solution to Datacenter Network Lag

admin — Mon, 21 Jul 2014 09:26:36 +0000

Next month, MIT researchers will present a breakthrough discovery that could change the way Web and mobile apps are written and help large corporations in making their datacentres more efficient.

Given that large websites have datacentres that are prone to congestion – “packets of data arriving at the same router at the same time are put in a queue, and if the queues get too long, packets can be delayed” – the new research has shown that the system can reduce network transmission queue length by over 99 percent.

In cooperation with Facebook, the MIT researchers experimented with one of the tech giants’ datacentres in an attempt to reduce the average queue length of routers. The research stated that, when traffic was most heavy, the average latency — the delay between the request for an item of information and its arrival – per request fell from 3.56 microseconds to .23 microseconds.

The model developed by the researchers – dubbed MIT Fastpass – replaces the standard decentralised networking model – where each node decides on its own when, where, and how to send data – to a centralized model called “arbiter” to decide “which nodes in the network may send data to which others during which periods of time.”

As ZDNet report, the research indicates:

“a single 8-core arbiter machine could handle 2.2 terabits of data per second, which, according to their announcement, equates to 2,000-gigabit connections running at full speed. The belief is that this could scale to a network of as many as 1,000 switches.”

It is believed that the Fastpass software will be released as open source, although the researches emphasised that it’s not a production-ready code. More information on the research will be presented at the ACM Special Interest Group on Data Communication this August.

Fujitsu Announce $2 Billion Investment in Cloud

admin — Tue, 15 Jul 2014 09:12:57 +0000

Japanese IT vendor Fujitsu announced on Monday that it is injecting $2 billion over the next two years to expand its cloud portfolio. The company said that the majority of its budget will be spent on new datacenter capacity and ramping up investment in infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS), software-as-a-service (SaaS), and cloud integration service, with the aim of accelerating its cloud sales to $3.45 billion.

“We see cloud as the natural platform for delivering these new types of applications, which is reflected in both the increase we are seeing in cloud adoption and how it is becoming a standard part of IT service delivery models for many CIOs today,” said Maeron McNaught, Fujitsu’s executive vice president of solutions and global delivery.

“We also recognize that no single cloud service provider can meet all customer needs, so we are continuing to invest in our cloud integration services,” he continued.

Currently, Fujitsu has 25 cloud centres globally and will deliver two additional data centers in the UK and US as part of the investment. ZDNet also report that Fujitsu have plans to open “two new locations in the USA for its IaaS Private Hosted deployments, plus additional deployments in six locations across Europe and Asia, as well as expansion in three existing geographies.”

Fujitsu is not alone in its huge investment in cloud; IBM and HP both announced this year plans to invest $1 billion in their respective cloud portfolios. However, Alexander Michael, an analyst at Frost & Sullivan, spoke of Fujitsu’s announcement favorably:

“Fujitsu has established itself in the global cloud computing market with its diverse portfolio and attractive cloud offerings, in spite of the high competition….“Given an end-to-end cloud services portfolio and unique value propositions, enterprises should definitely consider Fujitsu when evaluating cloud service providers.”