editor's blog
Subscribe Now

Tool Company Becomes Cloud Company

Continuing with our ongoing coverage of EDA in the public cloud, there’s another company that isn’t just starting to make use of the cloud: it’s completely tied its identity to the cloud in a way that entirely masks the original value proposition of the tools it makes.

The company is Nimbic.

Never heard of them?

How about PhysWare? Well, you might not have heard of them unless you’ve been involved in the world of field solvers. Which is what they do.  In fact, they claim to do it well – 10-100 times faster than their competition.

One of the things they’ve done well with this compute-intensive problem is to parallelize the algorithms, since the problem lends itself very nicely to concurrent programming. They can get a roughly 6x return by using 8 cores; not bad, given that Amdahl has limited how much benefit one can get.

Having done that, they took the next step: putting it into the cloud. Now users can marshall over a hundred Amazon machines to make quicker work of a thorny problem.

Of course, they’ve got their own way of doing things. While Synposys uses a command-line approach and Vector Fabrics uses a FLASH-based browser GUI, Nimbic has built their interface into their tool, which run on the local client (an approach Synopsys has indicated they will use at some point too). Bringing the cloud resources in to help accelerate a solution is a matter of a few menu clicks. They are considering a browser-based GUI, but using HTML5 instead of FLASH.

What happens behind the scenes is relatively straightforward. Subscribers have a small virtual machine that’s always running. When invoked, it will assemble a cluster of machines for the duration of an assignment. Actually, that’s not quite accurate: because machine billing happens on hour boundaries, if you end a job right after crossing an hour boundary, the cluster will stay alive for the rest of that hour – you’re paying for it anyway. That way if you need to do some more work, you don’t have to wait for the cluster to come up again.

And, as we’ve seen before, the cluster can take some time to be available – as much as 20 minutes for 100 machines (around 5 minutes for 5 machines).

With respect to security, all data is encrypted both in transmission and when stored. Neither Nimbic nor Amazon can read the data. While that data can remain in the machine for the duration of the account if desired, it’s not guaranteed – that is, it’s not intended as your repository. If the machine goes down for any reason, that data can be lost. So the expectation is that you have a client-side copy of everything, and the cloud storage is for convenience.

As to hacking protections, their VM has login disabled. They run a specific daemon, and only certain commands will be recognized. The “master” node communicates with the worker nodes securely by a locally-generated key, so not even Nimbic can snoop that. Even if someone managed to locate a machine and get something from it, this is a data-parallel problem: any machine only has a small portion of the overall design, making it even more difficult to piece together anything of value.

As with any cloud provider, they can provide regional containment in areas where Amazon has a center. So if a Japanese company can’t let their design leave Japanese soil, no problem: Amazon has a center in Japan. France? That’s tougher…

As to their value to their customers, they view it as either being peak resource alleviation (like Synopsys) for big companies, or an easier way for small and medium size companies to access the tools (like Vector Fabrics).

Looking forward, they see much more of this, although details are sketchy. After all, they’ve tied their entire company identity to the cloud. They see the possibility of offering more tools of their own through the cloud (not clear what those would be yet) or even hosting other companies’ tools (should they then call themselves CumuloNimbic?)

If they do the latter, they really stop being an EDA company (not a simple decision if they really do have a 10-100x advantage over their competition) and start being a tool cloud hosting company. So they clearly have some soul-searching to do as they decide what they want to be as they grow up.

You can find more info at their website.

Leave a Reply

featured blogs
Jun 30, 2020
“The customer is always right!” Anyone in customer service knows that motto. Pioneering businessmen César Ritz, Marshall Field, Harry Gordon Selfridge and others championed similar mantras throughout their careers. Today, Samtec lives and breathes Sudden Servic...
Jun 26, 2020
Oscar Vermeulen has been able to replicate the functionality of a classic PDP-11/70 using a mega-cool front panel with a Raspberry Pi as the computing engine....
Jun 26, 2020
[From the last episode: We looked at the common machine-vision application and its primary .] We'€™ve seen that vision is a common AI these days, and we'€™ve also talked about the fact that our current spate of neural networks are not neuromorphic '€“ that is, they'€™...

Featured Video

Product Update: Advances in DesignWare Die-to-Die PHY IP

Sponsored by Synopsys

Hear the latest about Synopsys' DesignWare Die-to-Die PHY IP for SerDes-based 112G USR/XSR and parallel-based HBI interfaces. The IP, available in advanced FinFET processes, addresses the power, bandwidth, and latency requirements of high-performance computing SoCs targeting hyperscale data center, AI, and networking applications.

Click here for more information about DesignWare Die-to-Die PHY IP Solutions

Featured Paper

Cryptography: A Closer Look at the Algorithms

Sponsored by Maxim Integrated

Get more details about how cryptographic algorithms are implemented and how an asymmetric key algorithm can be used to exchange a shared private key.

Click here to download the whitepaper

Featured Chalk Talk

Machine Learning at the Edge

Sponsored by Mouser Electronics and NXP

AI and neural networks are part of almost every new system design these days. But, most system designers are just coming to grips with the implications of designing AI into their systems. In this episode of Chalk Talk, Amelia Dalton chats with Anthony Huereca of NXP about the ins and outs of machine learning and the NXP iEQ Machine Learning Software Development Environment.

Click here for more information about NXP Semiconductors i.MX RT1060 EVK Evaluation Kit (MIMXRT1060-EVK)