Adaptiva OneSite and 1E SMSNomad are two innovative products, I always think why don’t Microsoft take over one of these products so that the clients of configmgr don’t want to pay more for the improved performance of a Microsoft product 🙂 Adaptiva Onesite Vs 1E Nomad.
I’ve another blog post which talks about the comparision of two products.
Through this blog I will try to document the comparision between both the configmgr (SCCM) extension products with the help of discussions on various forums.
Systems management across large global networks is complex and requires large investments in infrastructure. Adaptiva OneSite is a revolutionary add on to Microsoft’s ConfigMgr 2007 and SMS 2003 that enables you to manage a large distributed network as if it were a single local network.
– Roll-out ConfigMgr 2007 server infrastructure instantly: no additional hardware, do it in weeks
– Simplify operations: no system design or operational maintenance
– Lower costs: Reduce server cost, maximise bandwidth utilization
– Dramatically faster package downloads
– Secure software distribution and patching
– Fault tolerant and agile
Nomad Enterprise consists of two powerful components. It uses patented technology to optimize network efficiency and performance, enabling server consolidation and allowing customers to get the most out of their existing networks, avoiding costly network upgrades and the requirement for additional hardware.
– Nomad Branch® distributes systems management data once over the WAN and then shares it locally with peer agents in branch offices, with or without multicast. No physical visits to the branch sites are required.
– PXE Lite provides network booting capabilities without the need for separate server hardware, thereby reducing costs and making efficient use of network bandwidth and enabling real zero-touch bare-metal OS deployments.
I would like to add another good discussion which is happened on TechNet ConfigMgr forum (mainly, last part of the discussion)
Read more about the “Healthy Debate” and some main attractive parts (from my point of view) of discussion are highlighted below.
Comments from Deepak Kumar,
Chief Technology Officer, Adaptiva.
1. One Site works well with B/W accelerators out of the box. Our measurements are never fooled by them.
2. Version 1.2 of One Site will include a full Peer-to-peer PXE Server stack. You will not need to deploy anything separate like PXE Lite for OSD.
3. Master-slave architectures are obsolete for Peer-to-Peer because of the severe impact they cause on “Master” machines. One Site is based on a daisy chain architecture which is zero-impact for all participating peers. It is also more dynamic, allowing peers to come and go as they please, without resulting in any re-downloading of content.
You can download the recording of my detailed Webinar on the subject. It includes a presentation on the internal architecture of One Site.
If you’re evaluating OneSite against Nomad, please watch my detailed OneSite video on: http://adaptiva.com/videos.html
It includes an engineering presentation of our internal architecture and will greatly assist in your comparison.
1. Eliminating all DPs – Comments from mnikolajevs are correct. OneSite eliminates all DPs. You can use OneSite for small and large sites alike, not just branch offices, and remove all DPs from your environment. OneSite never needs a DP for any reason.
Nomad requires DPs for two reasons. Firstly, all master clients must connect to a DP to copy the package over. Second, every time any Nomad client, master or slave, copies any package, even using P2P, it must first go over the WAN to the DP and copy the LSZ file for that package from the DP. For Nomad to work, you must have a functioning DP infrastructure in place, with dependencies like site boundaries, distribution manager, and such.
2. Master – slave architecture vs. Diasy chaining – Adaptiva OneSite contains a modern zero-impact P2P design. All Adaptiva clients arrange themselves in a daisy chain. The first client gives content to client 2, which gives it to client 3, and so on. This means that each client serves packages to at most one other client at a time. It scales infinitely even in very large sites, because it has linear scalability charateristics. In a “Master-slave” architecture, multiple slaves download content from the same master. You can imagine what happens to the “Master” machine when multiple clients connect to it and simultaneously start copying files from it.
3. Adaptive bandwidth Management: Adaptiva OneSite contains an NDIS Kernel Driver that monitors the lengths of data queues at bottleneck routers in real time. It then places a single packet of data on the queue when it is nearly empty. As a result, we can harvest all UNUSED bandwidth, without competing with foreground traffic.
Nomad’s actual bandwidth management capabilities seem to be a bit ambiguous. According to a post from 1E’s Ed Aldrich, this is how it works:
Nomad client starts copying files from the DP across the WAN –
Every 4 seconds, Nomad client sends a ping across the WAN. If pings are taking longer to reach, it reduces the rate of file copying.
If it is true, this approach would be crude and ineffective, because by the time PING messages are delayed, the network has already been congested. The Adaptiva approach actually detects network congestion BEFORE it has happened, and prevents congestion, rather than reacting to it.
I’d love to see a clear explanation from 1E about Nomad’s actual bandwidth management capabilities, including definitions of the terms “Available bandwidth”, and “Work rate”.
4. Caching: Adaptiva OneSite contains a Kernel File System driver that combines the free disk space on all the computers in each office into a “Virtual SAN”. e.g. if there are 20 computers with 100 GB free space each, we’ll create a 2 TB “Virtual SAN” for you, and cache your packages on it as they arrive to the office. When any user needs more space, we detect their usage, and automatically release space to them without their even noticing it, since we are the file system. The benefit for you is that you get almost limitless cache space for free, and you’ll never need to manually manage this cache, or need to delete files from it.
Here, I would like to agree with Deepak’s on “Impact on Master“. I have personally experienced this problem for many branch offices in my previous assignment. It’s impacting the entire performance of the master. However, am not really sure, there are any changes in this on the latest releases of 1E SMSNomad.
See, the reply from Richard Threlkeld, 1E
Master-slave is obsolete? This is quite a naive statement, especially your point about the impact on the master. I’m sure our thousands of customers running Nomad on hundreds of thousands of systems would disagree with you. You also clearly do not know how Nomad works with it’s dynamic election process from your other statements as well.
Please don’t try to come onto a 1E forum and goad us into a debate by misrepresenting our software, especially when clearly your end result is to try and glean more information about how our software works. This frankly is unprofessional of you.
If there is curiosity about how Nomad works our customers will contact us thank you very much.
Personally, I don’t think these types of replies help 1E. Also, am curious to know why 1E is not explaining about their internal architecture (Did I miss anything?? 🙂 ).
Now, I would like share some neutral comments and clients feedback on both the products.
<Unfortunately, the actual comments removed from this blog as myITforum.com states that I need to get the “Consent” from them before puplishing their email comments in the blog> 🙁
Infact, Paul Thomsen was pointing out that the vendors shouldn’t be competing in community forums. The selection of the product should be based on “For What It’s Worth” !!!
Joel Grove, one of the very happy customer of Onesite had more detailed analysis and comparision report between these two products.