""Priscilla Oppenheimer"" wrote > I may be starting a new project doing some writing about > technologies used in enterprise networks. (read not service > provider) > > Do I need to cover IS-IS? Or is it mainly ISPs that use this?
I've never seen IS-IS in Enterprise networks, only ISP backbones and CODCN's (Central Office Data Communications Networks) that implement OSI stacks on network elements. The primary reason I've heard (I'm an IS-IS fan working in the Enterprise), that also happens to be a sort of compelling argument - is that OSPF and EIGRP work better through firewalls. I believe that IS-IS is significantly less complex than OSPF or EIGRP, and therefore, easier to train/learn, implement, manage, etc. > How about MPLS? I should discuss it briefly, but aren't > the main users of MPLS ISPs, not enterprise networks? I've only seen MPLS in ISP networks, but could be used in Enterprise, especially simple VRF concepts (from MPLS VPN's). I could see many reasons that Enterprises could use even static route VRF's for various useful purposes on occasion, but even that is a rarity. Of course, I'm also of the opinion that MPLS doesn't really belong in SP networks either - so YMMV. MPLS-TE is still being argued among the SP network guru's. MPLS FRR (Fast Re-Route) is an important network protection and resliency technology and should be researched, tested, and implemented when the need arises for it (including for Enterprises). > Anyone using GARP? That's on my list to research too. GARP, as in Multicast? Many Enterprises are using Multicast, but their use can vary depending on the company (could be for reliable multicast like TIBCO, or could be for live broadcasting sending/receiving, or research, or many other technical/business reasons). Understanding the basics is good, but it's hard to cover all the advanced topics in IP Multicast because it's such a point application as it is already. > Alas, I have a lot to learn. Thank-you VERY much > for answering these quick questions. You will probably want to look at this from a specific industry perspective, i.e.: High-Tech Manufacturing / Hardware/Software Engineering Pharmaceuticals Automotive, Consumer Goods Government / Political / Federal vs. State/County/City Defense / Military / Aerospace Education Banking / Securities / Financials / Real-Estate Insurance Companies Entertainment Retail Health Care Hospitality / Transportation Energy Certain companies have totally different needs technology-wise... in particular, some business build networks without latency in mind because they are a state-wide organization, or even a city-wide organization. Some companies have networks that span the globe, and some have the same type of need in a mission-critical way. There are many technologies that apply to certain companies and not others. For any large Enterprise that relies on IT to be mission-critical, figure that the cost of IT downtime is significant (Meta Group 2001 report shows $1.5M per hour of downtime). IT relies on the Internet for growth and technology. Data centers, especially Internet data centers are the key point to any IT infrastructure. The primary technologies in these environments are rapidly shifting. Many organizations are now required to, or at the very least are planning to, add more than one primary data center for 100%, near-instantaneous continuous operation / disaster recovery. You will see this in the High-Tech industry because of maturity, and you will see it in the Government/Defense industries because of things like the Homeland Security Act (signed November 25th, 2002). The same Meta group report shows that 70-75% of mid- to top-tier applications will be distribued across at least two data centers by 2003. I got this information from: http://www.cisco.com/global/EMEA/networkers/presentations/NSC-284_Reiner_Dre sbach.pdf.gz Because of the need for distributed data centers, technologies such as: BGP-4 (and therefore IP address management/globalization, as well as Peering and Carrier-Neutral co-location) Content Networking (RHI, E-CDN, DNS/HTTP-routing) Optical Networking (DWDM, Optical Ethernet, RSTP) Virtual Private Networking (to include Site-to-Site VPN, Remote Access User VPN - both over the Internet and on private networks, as well as CE MPLS VPN) Security (I see this more as host security, then network security, but YMMV) will become increasingly important for Enterprises. More importantly, they will want to increase availability and performance of 3-tier applications (front-ends, application servers, and database servers), while also making them global and distributed. This will have a very serious impact on the networking technologies and industry best practice methods. Also big for Enterprises will be technologies that represent significant ROI with quick and easy Break-Even points, etc. i.e.: IP Telephony (SIP, Unified Messaging, CallManger, Unity) Network Management / OSS / Automation 3-Tier Applications are really big in large Enterprises, and there are only a few current methods of implementing them. IBM - pSeries (AIX/Linux) + Websphere + DB2 Sun - Ultra (Solaris) + BEA WebLogic + Oracle 9i Microsoft - HP or Dell (NT server) + .NET + MS SQL Some people do multi-vendor mix 'n match, but that is the general idea. What's even funnier is look at this model: Largest Enterprise (Fortune 500) -> IBM Other Large Enterprises -> Best Practice (Sun+BEA+ORA) Enterprise businesses -> Cheapest (Microsoft+Dell) The nice thing about being a Network engineer is that you are often stuck with Cisco (only other choices are really Juniper, Foundry, Extreme, Riverstone - who don't have solutions, only point products), but any basic Cisco network will nicely support any of the three most popular 3-tier application niches (or any variant). The network requirements don't change, so it's much easier to sell/implement/fund/support/etc. Do a search on CCO (logged in with a customer account) for "SRND" and read all those PDF's. You will see what an Enteprise might need today. In reality, these technologies are going to represent significant cost savings for Enterprises, however, they will have to consider that more spending will occur in the short-term (CAPEX often leads to lower future OPEX), but many businesses are especially thinking about long-term viability in this current economy (funny how things work that way). The ROI on IP Telephony alone can represent such significant savings, it's very logical for Enterprises to move to all sorts of complex networks that require new business units to build and support (yet that still manage to consistently meet break even points, etc). What is it that you are working on? - dre Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=58506&t=58493 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
