The Lustre HPC file system has another new home — and it’s the right one

After years of bouncing from one company to another, the Lustre file system that is so popular in high-performance computing (HPC) has been sold yet again. This time it went to an enterprise storage vendor. Finally, it’s in the hands of a company that makes sense.DataDirect Networks (DDN) announced it purchased the Lustre File System from Intel this week as Intel looks to pare down non-essential products. DDN got all assets and the Lustre development team, who are undoubtedly relieved. The announcement was made at the International Supercomputing Conference (ISC) in Frankfurt, Germany.[ Check out AI boosts data-center availability, efficiency. Also learn what hyperconvergence is and whether you’re ready for hyperconverged storage. | For regularly scheduled insights, sign up for Network World newsletters. ] Lustre's history Lustre (which is a portmanteau of Linux and cluster) is a parallel distributed file system that supports multiple computer clusters with thousands of nodes. It started out as an academic research project and was later acquired by Sun Microsystems, which was in turn acquired by Oracle.To read this article in full, please click here

Cryptocurrency API Gateway using Typescript+Workers

If you followed part one, I have an environment setup where I can write Typescript with tests and deploy to the Cloudflare Edge with npm run upload. For this post, I want to take one of the Worker Recipes further.

I'm going to build a mini HTTP request routing and handling framework, then use it to build a gateway to multiple cryptocurrency API providers. My point here is that in a single file, with no dependencies, you can quickly build pretty sophisticated logic and deploy fast and easily to the Edge. Furthermore, using modern Typescript with async/await and the rich type structure, you also write clean, async code.

OK, here we go...

My API will look like this:

Verb Path Description
GET /api/ping Check the Worker is up
GET /api/all/spot/:symbol Aggregate the responses from all our configured gateways
GET /api/race/spot/:symbol Return the response of the provider who responds fastest
GET /api/direct/:exchange/spot/:symbol Pass through the request to the gateway. E.g. gdax or bitfinex

The Framework

OK, this is Typescript, I get interfaces and I'm going to use them. Here's my ultra-mini-http-routing framework definition:

export interface IRouter {
  route(req: RequestContextBase): IRouteHandler;
}

/**
 * A route
 */
export interface IRoute  Continue reading

List of EVPN and DC-related IETF drafts (as of 1H 2018)

Ethernet Virtual Private Network (EVPN) solution is becoming pervasive for Network Virtualization Overlay (NVO) services in data center (DC) networks and as the next generation VPN services in service provider (SP) networks. As a result of this popularity a lot of work is going on in IETF in this area. In this post I collect links to some of the interesting IETF drafts in this area.
All this information is relevant for 1H 2018 timeframe. Keep in mind to pay attention to which version of draft you’re reading – this drafts are frequently updated.

 

Service Chaining using Virtual Networks with BGP VPNs

This document describes how service function chains (SFC) can be applied to traffic flows using routing in a virtual (overlay) network to steer traffic between service nodes. Chains can include services running in routers, on physical appliances or in virtual machines. Two techniques are described: in one the service chain is implemented as a sequence of distinct VPNs between sets of service nodes that apply each service function; in the other, the routes within a VPN are modified through the use of special route targets and modified next-hop resolution to achieve the desired result.

 

Applicability of Continue reading

Want affordable SDN products? Check the secondary market

Frankly (no pun intended), I have to admit that I’m growing increasingly frustrated with certain trends in networking.For example, it’s not that I don’t like the dream or idea of software-defined networking (SDN) — it’s not that I don’t think it’s superior to the older way of setting up or monitoring a network. It’s just that I’m becoming increasingly concerned that small- to medium-size enterprises (SMEs) won’t be able to keep up. And the media that follows this trend isn’t really brining to light the extreme cost of some of these systems.Pricewise, many of the product lines are intended for large networks. There's no way that a smaller company could even begin to afford them. For example, one trainer told me that a certain SDN product was scaled to start at 500 site deployments!!To read this article in full, please click here

Want affordable SDN products? Check the secondary market

Frankly (no pun intended), I have to admit that I’m growing increasingly frustrated with certain trends in networking.For example, it’s not that I don’t like the dream or idea of software-defined networking (SDN) — it’s not that I don’t think it’s superior to the older way of setting up or monitoring a network. It’s just that I’m becoming increasingly concerned that small- to medium-size enterprises (SMEs) won’t be able to keep up. And the media that follows this trend isn’t really brining to light the extreme cost of some of these systems.Pricewise, many of the product lines are intended for large networks. There's no way that a smaller company could even begin to afford them. For example, one trainer told me that a certain SDN product was scaled to start at 500 site deployments!!To read this article in full, please click here

Google cloud storage gets a boost with managed NAS service

Google is adding to its cloud storage portfolio with the debut of a network attached storage (NAS) service.Google Cloud Filestore is managed file storage for applications that require a file system interface and shared file system for data. It lets users stand up managed NAS with their Google Compute Engine and Kubernetes Engine instances, promising high throughput, low latency and high IOPS.[ Check out AI boosts data-center availability, efficiency. Also learn what hyperconvergence is and whether you’re ready for hyperconverged storage. | For regularly scheduled insights sign up for Network World newsletters. ] The managed NAS option brings file storage capabilities to Google Cloud Platform for the first time. Google’s cloud storage portfolio already includes Persistent Disk, a network-attached block storage service, and Google Cloud Storage, a distributed system for object storage. Cloud Filestore fills the need for file workloads, says Dominic Preuss, director of product management at Google Cloud.To read this article in full, please click here

MTU and TCP MSS

The largest chunk of bytes that a transport protocol can forward across specific medium is called MTU – Maximum Transmission Unit. If we speak about Ethernet, which is today the most common, he has by default 1522 bytes MTU. The story about MTU is that the MTU of specific protocol basically defines how much payload (or highest protocols headers + their payload) it can carry in its biggest packet, not counting his own headers. Putting more payload into single packet than the MTU allows will result in fragmentation, the process of slicing the frame into more smaller frames so that

The post MTU and TCP MSS appeared first on How Does Internet Work.

Debugging data flows in reactive programs

Debugging data flows in reactive programs Banken et al., ICSE’18

To round off our look at papers from ICSE, here’s a really interesting look at the challenges of debugging reactive applications (with a certain Erik Meijer credited among the authors).

… in recent years the use of Reactive Programming (RP) has exploded. Languages such as Elm and libraries such as Reactor, Akka, and Rx are being used by companies such as Netflix, Microsoft, and Google, to build highly responsive and scalable systems.

The rise of reactive programming fits well with the increasing need to process streams of data. In a reactive program, you set up a data processing pipeline and then wait for input to arrive.

Many RP implementations share a notion of a collection that abstracts over time, in contrast to space like standard collections. This collection comes in different flavors, such as Observable (Rx)… the implementations differ in the precise semantics of their collections, their execution model (push/pull) , and the set of available operators.

While in theory events and threads are duals, in practice the RP abstraction works very well for expressing streaming pipelines. If you have an abstraction over time though, Continue reading

Viptela SDWAN Overview

Viptela is an SDWAN company that was recently purchased by Cisco. In this post I will go over the components that make up the Viptela secure extensible network (SEN) solution. By the end of this post you will have a good idea of the pieces that make up the Viptela puzzle. Network...

OpenStack Summit Vancouver recap: What we did and what’s next

A few weeks ago, I set out to the beautiful city of Vancouver’s convention center, along with a boatload of rocket turtles and a stellar team. It was a great time with a wonderful scenic view of the ocean. I’ve been at Cumulus a few months now, but I can’t help but enjoy looking around, seeing the friends I’ve made in the industry, and the friends and companies Cumulus has worked with over the years. It is exciting to have thousands of people coming together at OpenStack Summit Vancouver to work on a shared goal.

OpenStack Summit Vancouver

This year, we were lucky enough to have our very own Pete Lumbis take the stage with David Iles of Mellanox to present our joint solution around the latest SDN revolution, which is centered on creating efficient virtualized data center networks using VXLAN & EVPN.

In the next few paragraphs, I’ll share some highlights of the event, some photos, and a recap of that exact discussion. There was a lot to learn and discover, and I’m excited to share the details.

Day 1: Don’t you just love working with collaborative open source projects?

On our first day, lots of things were going on — we Continue reading

Train Your Team With INE And Reap The Benefits Of Becoming A Cisco Partner

Training with INE will not only keep you team on the cutting edge of the industry’s leading networking technologies, it will also put you on the path to becoming a Cisco Partner; a tangible way to show clients that you’re on the forefront of Cisco technologies and training. Cisco registered partners enjoy access to a wide variety of Cisco channel partner tools and are eligible to apply for Cisco certifications and specializations.

Additional Benefits Include:
  • Credibility: Capitalize on the Cisco Brand, which commands recognition and respect in the marketplace.
  • Credentials: Market your certified Cisco Expertise as a competitive differentiator to customers.
  • Momentum: Join the value-add partner program that is setting the pace in the industry.
  • Rewards: Add value for customers and drive growth and profitability with Cisco.
  • Relationships: Position your company to take advantage of more sales opportunities than ever before.

There are three levels of certification that a team can achieve; Cisco Select, Premier and Gold. Each level requires certain prerequisites, which INE’s industry-leading network training can provide. The three types of certifications are:




For more information contact our training advisors at [email protected] or by phone at +1-877-224-8987 or +1-775-862-4344 (outside the U.S.)

IDG Contributor Network: Skimping on business data protection could be a costly mistake

Data is essential to the smooth operation of any organization. Whether it’s data on your products, customers, or competition, you need it to do business. Your software and systems are dependent on the data that’s fed into them.Big data may be gathered by IoT sensors in vehicles and buildings, smartphones, and from countless other data points to inform big decisions. But at a granular level you also need small pieces of data to function. Without credentials you can’t gain access to the big data, contact suppliers, or even tweak the air conditioning system.Our dependence on data is profound, you might say it’s your business DNA because it’s crucial for survival and growth.To read this article in full, please click here

1 … 1,508 1,509 1,510 1,511 1,512 … 3,798